A monitoring and control method, system, medium and program product for regional electricity consumption

Abstract

A monitoring and control method, system, medium and program product for regional electricity consumption, which relates to the field of terminal user power distribution system, and includes: constructing a multi-level electricity consumption model based on regional power network structure data and user power consumption settings; obtaining the associated characteristic data of the target room including the number of people in the room and the equipment list, and determining the real-time business scenario of the target room; determining the electricity consumption data boundary of the target room based on the historical electricity consumption data of multiple similar rooms in the room layer; when the real-time electricity consumption data does not meet the electricity consumption data boundary, the target room is regarded as an abnormal room, and the abnormal equipment and abnormal floor corresponding to the abnormal room are determined; abnormal characteristics are determined according to the real-time business scenario and real-time electricity consumption data, and equipment optimization instructions are generated; based on the proportion of electricity consumption of abnormal rooms in abnormal floors, the power supply scheduling of multiple room layers in abnormal floors is controlled. Implementation of this method can improve electricity efficiency, optimize power resource allocation, and thus ensure electricity safety.

Description

Technical Field

[0001] The present application relates to the field of terminal user power distribution systems, and in particular to a monitoring and control method, system, medium and program product for regional power consumption. Background Art

[0002] With the accelerated advancement of industrialization and urbanization, the management and optimization of power resources have become particularly important. In commercial or residential buildings such as schools and hospitals, effective power monitoring can not only ensure power supply safety, but also improve energy efficiency and reduce costs. Therefore, the development of an efficient power monitoring system is an important direction of current technological development.

[0003] In the related art, the power consumption monitoring system mainly relies on a central control unit to process the collected data, and organizes and generates reports on the power consumption data through a unified management software. Although such a system can process a large amount of data, it mainly focuses on monitoring the total power consumption, and rarely involves monitoring the specific power consumption of a single room or device.

[0004] The current single-level monitoring system is often difficult to be effectively applied when facing complex power usage scenarios with multiple devices and multiple rooms. Due to the lack of real-time monitoring and analysis of the specific power usage of each room or device, the system often finds it difficult to quickly locate the source of the problem when power usage anomalies occur, and it is also difficult to effectively allocate power according to the actual power demand of different rooms or devices, resulting in energy waste and safety risks. Summary of the invention

[0005] The present application provides a monitoring and control method, system, medium and program product for regional electricity consumption, which are used to model and divide the monitoring area, realize real-time monitoring and analysis of multiple devices, improve electricity efficiency, optimize power resource allocation, and thus ensure electricity safety.

[0006] In the first aspect, the present application provides a method for monitoring and controlling regional electricity consumption, which is applied to an electricity consumption monitoring system, and the method comprises: constructing a multi-level electricity consumption model including a device layer, a room layer and a building layer based on regional power network structure data and user electricity consumption settings; there are multiple floors in the building layer, multiple rooms in the room layer, and multiple devices in the equipment layer, one floor corresponds to multiple rooms, and one room corresponds to multiple devices; obtaining associated feature data of the target room including the number of people in the room and the equipment list, and determining the real-time business scenario of the target room based on the associated feature data; the target room is a room in the room layer; based on the historical electricity consumption data of multiple similar rooms in the room layer on the multi-level electricity consumption model , determine the power consumption data boundary of the target room in the real-time business scenario; the equipment corresponding to the target room is the same as that of the same type of room; when it is detected that the real-time power consumption data corresponding to the target room exceeds the power consumption data boundary, the target room is regarded as an abnormal room, and the abnormal equipment and abnormal floor corresponding to the abnormal room are determined based on the multi-level power consumption model; after determining the abnormal characteristics of the abnormal equipment according to the real-time business scenario and the real-time power consumption data, generate equipment optimization instructions based on the abnormal characteristics and send the equipment optimization instructions to the abnormal equipment; the equipment optimization instructions include adjusting operating parameters or temporarily cutting off power supply; based on the abnormal characteristics and the proportion of power consumption of the abnormal room in the abnormal floor, control the power supply scheduling of multiple room layers in the abnormal floor.

[0007] In the above embodiment, the electricity monitoring system can identify the business scenarios of the room in real time based on the associated characteristic data such as the number of people in the room and the equipment list, and adjust the electricity usage strategy accordingly. This dynamic adjustment mechanism greatly improves the accuracy and efficiency of energy use. When abnormal electricity consumption data is detected, the system can quickly locate the abnormal room and equipment, respond quickly and make adjustments, thereby reducing safety risks and reducing energy waste. In addition, by intelligently generating equipment optimization instructions, such as operating parameter adjustments or temporary power cut-off, not only the equipment operation efficiency is improved, but also the safety of electricity use is guaranteed. Through its multi-level detailed monitoring and intelligent response mechanism, the management efficiency and safety of electricity consumption in the building are significantly improved, while also bringing energy savings and cost reductions.

[0008] In combination with some embodiments of the first aspect, in some embodiments, associated feature data of the target room including the number of people in the room and the equipment list is obtained, and the real-time business scenario of the target room is determined based on the associated feature data, specifically including: obtaining a real-time image of the target room, and determining the number of people in the room based on the real-time image; determining the equipment list of the target room in combination with the real-time image and user settings; the equipment list includes device model information, power parameters and quantity information of various electrical equipment configured in the target room; taking the number of people in the room and the equipment list as associated feature data, determining a business adaptation set according to the target room and the associated feature data; inputting the associated feature data into a business scenario recognition model corresponding to the business adaptation set to obtain a real-time business scenario; when the number of people in the room or the equipment list changes, re-obtaining the changed associated feature data, and dynamically updating the real-time business scenario of the target room.

[0009] In the above embodiment, the electricity consumption monitoring system first obtains the real-time image of the target room through the camera, and determines the number of people in the room based on the real-time image. At the same time, the equipment list of the target room is determined in combination with the real-time image and user settings. The number of people in the room and the equipment list are used as associated feature data, and the business adaptation set is determined according to the target room and the associated feature data. The associated feature data is input into the business scene recognition model to obtain the real-time business scene. When the number of people in the room or the equipment list changes, the system will also re-acquire the changed associated feature data and dynamically update the real-time business scene of the target room. By integrating image information and user settings, the real-time status data of the room is fully obtained, and then combined with business adaptation and scene recognition, the actual business scene of the room can be accurately judged. At the same time, it supports dynamic updating of scenes, so that the system can adapt to real-time changes in the room usage status and improve the accuracy of electricity consumption monitoring.

[0010] In combination with some embodiments of the first aspect, in some embodiments, the number of people in the room and the equipment list are used as associated feature data, and a service adaptation set is determined according to the target room and the associated feature data, specifically including: presetting multiple corresponding adaptation service scenarios according to the room type and room purpose of the target room; in the adaptation service scenario, filtering out a first service subset with a number of people above a first threshold according to the number of people in the room; obtaining a historical equipment list of the room, comparing the historical equipment list with the equipment list, and obtaining a newly added equipment list and a reduced equipment list; if both the newly added equipment list and the reduced equipment list are empty, obtaining the historical service scenarios of the target room at the previous moment; scene, and according to the historical business scenarios, a second business subset whose deviation from the historical business scenarios is less than a second threshold is screened out from the first business subset; if the newly added equipment list is not empty, the equipment in the newly added equipment list is matched with the standard equipment corresponding to each business scenario in the first business subset, and according to the matching degree, a second business subset whose matching degree is greater than a third threshold is screened out from the first business subset; if the reduced equipment list is not empty, the business scenarios containing the equipment in the reduced equipment list in the standard equipment corresponding to each business scenario in the first business subset are removed to obtain the second business subset; and the second business subset is determined as the business adaptation set of the target room.

[0011] In the above embodiment, the power consumption monitoring system presets multiple adaptation business scenarios according to the room type and purpose of the target room. Then, from the adaptation business scenarios, the first business subset with a high number of people is screened out according to the number of people in the room. Then, the historical equipment list of the room is compared with the current equipment list to obtain a list of newly added equipment and a list of reduced equipment. If both the newly added and reduced equipment lists are empty, a second business subset with a small deviation from the historical business scenario is screened out from the first business subset according to the historical business scenario of the target room; if the newly added equipment list is not empty, the matching degree between the newly added equipment and each business scenario in the first business subset is calculated, and a second business subset with a large matching degree is screened out; if the reduced equipment list is not empty, the business scenario containing the reduced equipment in the first business subset is removed to obtain the second business subset; and finally the second business subset is determined as the business adaptation set of the target room. Taking into account multiple factors such as room attributes, changes in the number of people and changes in equipment, adaptive matching of room business scenarios is achieved, misjudgment is avoided, and the reliability of real-time business scenario recognition is improved.

[0012] In combination with some embodiments of the first aspect, in some embodiments, after determining the abnormal characteristics of the abnormal device based on the real-time business scenario and real-time power consumption data, generate a device optimization instruction based on the abnormal characteristics and send the device optimization instruction to the abnormal device, specifically including: determining the various power consumption parameters of the abnormal device in multiple power consumption time periods according to the real-time business scenario and real-time power consumption data; the power consumption parameters include voltage, current, power factor and harmonics; determining the abnormal characteristics of the abnormal device including the degree of equipment performance degradation, power consumption risk level and energy efficiency level according to the various power consumption parameters; according to the abnormal characteristics, matching the device optimization strategy in the preset strategy library, and generating the corresponding device optimization instruction based on the device optimization strategy; the device optimization instruction includes one or more of the following: adjusting the operating parameters of the device, updating the control logic of the device or cutting off the power supply of the device; sending the device optimization instruction to the control module of the abnormal device, so that the abnormal device executes the corresponding optimization instruction.

[0013] In the above embodiment, the power consumption monitoring system first determines the various power consumption parameters of abnormal equipment in multiple power consumption periods based on real-time business scenarios and real-time power consumption data. Based on these power consumption parameters, the abnormal characteristics of the abnormal equipment are determined, including the degree of equipment performance degradation, power consumption risk level and energy efficiency level. According to the abnormal characteristics, the corresponding equipment optimization strategy is matched in the preset strategy library, and the equipment optimization instructions are generated accordingly. The instructions include adjusting the equipment operating parameters, updating the control logic or cutting off the power supply. Finally, the optimization instructions are sent to the control module of the abnormal equipment to guide the equipment to perform the corresponding optimization actions. By comprehensively collecting the multi-dimensional power consumption parameters of the equipment, determining its inherent abnormal characteristics, and then corresponding to the preset optimization strategy, practical optimization instructions are automatically generated, forming a complete equipment abnormality monitoring and optimization mechanism. It not only improves the efficiency of discovering and handling equipment anomalies, but also ensures power safety and energy efficiency.

[0014] In combination with some embodiments of the first aspect, in some embodiments, the multi-level electricity consumption model also includes a building layer and a regional layer. According to the abnormal characteristics, the device optimization strategy is matched in the preset strategy library, and the corresponding device optimization instructions are generated based on the device optimization strategy, specifically including: in the preset strategy library, according to the abnormal characteristics, the first optimization strategy for device layer abnormalities, the second optimization strategy for room layer abnormalities, the third optimization strategy for floor layer abnormalities, the fourth optimization strategy for building layer abnormalities and the fifth optimization strategy for regional layer abnormalities are matched; the first optimization strategy is used to instruct to adjust the operating parameters of the abnormal device or temporarily cut off its power supply; the second optimization strategy The first optimization strategy is used to indicate power limitation or timed power-off of abnormal equipment in the room layer; the third optimization strategy is used to indicate total control or time-sharing power supply to abnormal rooms in the building layer; the fourth optimization strategy is used to indicate power load balancing between floors in the building layer; the fifth optimization strategy is used to indicate power dispatching for buildings in the regional layer; according to the abnormal level corresponding to the abnormal characteristics, the target optimization strategy is determined among the first optimization strategy, the second optimization strategy, the third optimization strategy, the fourth optimization strategy and the fifth optimization strategy; according to the target optimization strategy, the equipment optimization instructions of the corresponding level are generated, and the equipment optimization instructions are sent to the node devices of the corresponding level.

[0015] In the above embodiment, the electricity monitoring system predefines multiple optimization strategies for different abnormal levels in the preset strategy library. These hierarchical optimization strategies cover various ranges from local to global. For example, the device layer strategy focuses on the adjustment of single device parameters, while the regional layer strategy focuses on power dispatch between building complexes. According to the abnormal level to which the abnormal characteristics belong, the system selects the corresponding target optimization strategy, and generates equipment optimization instructions for this level based on the strategy, which are sent to the node devices of the corresponding level. By building a multi-level strategy library, exclusive optimization strategies are customized for abnormal situations of different granularity and range, and the corresponding execution subjects are matched. This hierarchical and graded abnormal handling mechanism can more accurately and efficiently deal with various types of electricity anomalies in complex regional environments, give play to the local and overall synergistic optimization effects, and maximize the regional electricity safety and efficiency.

[0016] In combination with some embodiments of the first aspect, in some embodiments, based on abnormal characteristics and the proportion of electricity consumption of abnormal rooms in abnormal floors, the power supply scheduling of multiple room layers in abnormal floors is controlled, specifically including: obtaining a power supply line topology map of the abnormal floor, and identifying other rooms located on the same power supply branch as the abnormal room based on the power supply line topology map; calculating the proportion of the power load of the abnormal room to the total load of the abnormal floor; when the power consumption proportion exceeds a preset power consumption threshold, short-term power limitation of other rooms to ensure power supply to the abnormal room.

[0017] In the above embodiment, the power consumption monitoring system first identifies other rooms located on the same power supply branch as the abnormal room based on the topological map. Then the proportion of the power load of the abnormal room to the total load of the entire abnormal floor is calculated. When the ratio exceeds the preset threshold, short-term power restrictions are implemented on other rooms on the same branch as the abnormal room to ensure power supply to the abnormal room. By analyzing the power supply topology of the abnormal floor, other rooms associated with the abnormal room are accurately located, and the abnormal room is taken as the key protection object, while ensuring the power supply of the abnormal room, the impact on other areas of the floor is reduced. This refined multi-room collaborative control mechanism based on the internal topological association of the floor can take into account the overall power supply quality when dealing with local anomalies, thereby ensuring power safety.

[0018] In combination with some embodiments of the first aspect, in some embodiments, based on regional power network structure data and user power settings, a multi-level power consumption model including device layer, room layer and building layer is constructed, specifically including: receiving and parsing regional power network structure data, determining the device nodes, room nodes, floor nodes and the topological connection relationship between nodes in the regional power network structure data; accessing the power consumption information database, and extracting the attribute information of each node; the attribute information includes the rated parameters of the equipment, the area of ​​the room, the functional attributes and the vertical position information of the floor; generating a multi-level power consumption topology network according to the topological connection relationship and the attribute information; the multi-level power consumption topology network is a hierarchical structure including device layer, room layer and building layer; receiving the power consumption settings entered by the user in the user terminal, and mapping and labeling the multi-level power consumption topology network with the power consumption settings as constraints; the power consumption settings include the power consumption time period of each room, the rated power of each device and the power consumption priority parameters; based on the labeled multi-level power consumption topology network, constructing a multi-level power consumption model.

[0019] In the above embodiment, the power consumption monitoring system first receives and parses the regional power network structure data to determine the topological connection relationship between the device nodes, room nodes, floor nodes and nodes. Then access the power consumption information database to extract the attribute information of each node, including the rated parameters of the equipment, the area and functional attributes of the room, the vertical position of the floor, etc. Based on the topological relationship and attribute information, a multi-level power consumption topological network including the device layer, the room layer and the floor layer is generated. On this basis, the power consumption time period of each room, the rated power and priority parameters of each device set by the user in the terminal are received as constraints, and the multi-level power consumption topological network is mapped and annotated, and finally a multi-level power consumption model is constructed. By directly using the regional power grid structure data, the multi-level topological relationship can be extracted quickly and accurately. By integrating the attributes of the equipment, room, and floor in the database and superimposing the constraints set by the user, the multi-level power consumption model can comprehensively and finely characterize the static structure and dynamic demand of the entire area, ensuring the normal progress of subsequent power consumption abnormality analysis, multi-dimensional monitoring and hierarchical control.

[0020] In a second aspect, an embodiment of the present application provides an electricity usage monitoring system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the electricity usage monitoring system to execute the method described in the first aspect and any possible implementation method of the first aspect.

[0021] In a third aspect, an embodiment of the present application provides a computer program product comprising instructions, and when the above-mentioned computer program product runs on an electricity consumption monitoring system, the above-mentioned electricity consumption monitoring system executes the method described in the first aspect and any possible implementation method of the first aspect.

[0022] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, comprising instructions. When the instructions are executed on a power consumption monitoring system, the power consumption monitoring system executes the method described in the first aspect and any possible implementation method of the first aspect.

[0023] It is understandable that the power consumption monitoring system provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the method provided in the embodiment of the present application. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method, which will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0025] 1. Due to the use of multi-level power consumption monitoring and real-time business scenario identification technology, it is possible to accurately adjust power consumption strategies and optimize power supply scheduling, effectively solving the problems of extensive power consumption management and inefficient energy distribution in related technologies, thereby achieving improved power safety and reduced energy consumption. By subdividing power consumption monitoring into equipment layer, room layer and building layer, each level can perform independent power consumption analysis and regulation based on real-time data. At the same time, combined with real-time business scenario identification, the system can adjust power consumption strategies in real time according to dynamic information such as changes in the number of people in the room and equipment usage status, such as reasonably allocating loads and optimizing power supply time periods, so as to more accurately meet actual power demand and reduce energy waste. It not only improves the response speed and safety of the power consumption system, but also improves energy efficiency.

[0026] 2. Due to the adoption of an integrated abnormal handling process driven by equipment power consumption parameter extraction, abnormal feature analysis and strategy matching, it is possible to generate accurate and feasible optimization instructions for a single abnormal device, effectively solving the problem of rough and general optimization measures for abnormal equipment in related technologies, and thus achieving efficient closed-loop handling of equipment-level abnormalities. The optimization measures obtained by matching are more professional and sophisticated, forming a closed loop from abnormality discovery to abnormality handling. It not only effectively avoids the risk of electrical accidents, but also extends the service life of equipment, and continuously optimizes energy conservation and consumption reduction, to ensure safe electricity use in the region.

[0027] 3. Due to the use of regional power supply line topology analysis and abnormal room association identification, the potential chain effects caused by abnormal rooms can be discovered, effectively solving the problem of local abnormal handling ignoring the overall power supply quality in related technologies, thereby achieving balanced and reliable power supply for floors and even the entire building. When dealing with abnormalities in a single room, it is possible to predict the spatial scope of its impact and identify other affected rooms. Through differentiated power regulation, while ensuring that abnormal rooms are given priority power supply, the negative impact is controlled to the minimum range. This collaborative mechanism greatly improves the overall power supply safety index of the building and ensures the stable operation of key places. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a flow chart of a method for monitoring and controlling regional electricity consumption in an embodiment of the present application;

[0029] Figure 2 is another flow chart of the method for monitoring and controlling regional electricity consumption in an embodiment of the present application;

[0030] Figure 3 It is a schematic diagram of the structure of a physical device of the electricity consumption monitoring system in the embodiment of the present application. DETAILED DESCRIPTION

[0031] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification of the present application, the singular expressions "one", "a kind of", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations comprising one or more of the listed items.

[0032] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0033] For ease of understanding, the application scenarios of the embodiments of the present application are introduced below.

[0034] As a regional medical center, a tertiary hospital has a building area of ​​200,000 square meters, an average daily outpatient volume of more than 5,000 people, and more than 50 clinical and medical technology departments. With the continuous improvement of the intelligence and information level of medical equipment, the power load within the hospital is increasing, especially during peak hours, there is a great conflict in the power demand of various departments, and power shortages and even power outages often occur in some areas.

[0035] In related technologies, a central power monitoring system can be used to centrally monitor total power current, voltage and other parameters in the floor distribution room, and set safety thresholds to achieve protective control of the power load of the entire floor. However, this technology is difficult to finely control and dispatch the power consumption of each department, and it is difficult to deal with the chain reaction caused by local anomalies, which poses a great safety hazard.

[0036] The regional electricity consumption monitoring and control method in the embodiment of the present application is adopted. By constructing an electricity consumption model including multi-dimensional topological relationships of equipment, rooms, and floors, real-time status data such as room traffic and equipment lists are obtained, and room-level business scenario changes are dynamically perceived. Based on this, accurate room-level abnormal electricity consumption warnings are generated, thereby achieving refined monitoring and hierarchical management of electricity consumption behavior within the hospital. It can not only accurately locate and quickly deal with local electricity consumption anomalies and avoid accident risks, but also coordinate redundant power resources. When key rooms are overloaded, intelligent peak shaving is implemented for non-core loads, thereby ensuring continuous power supply to key areas without affecting the overall power supply quality.

[0037] The following describes a scenario in which the method for monitoring and controlling regional electricity consumption in this application is used.

[0038] After the hospital deployed the power consumption monitoring system of this application, the system first automatically generated a multi-level power consumption topology model for the entire hospital through image recognition and other technologies, which can clearly present the spatial location relationship and power supply connection relationship of different dimensions such as hospital area, building, floor, department, room, etc. In daily operation, the system collects data such as the flow of people and equipment list in each room in real time, and dynamically evaluates the real-time power consumption risk level of each room in combination with the room business attributes and equipment rated parameters.

[0039] One day, when the air conditioner, electric blanket, suction device, fetal heart monitor and other equipment in the obstetrics and gynecology delivery room were running at the same time, the system promptly discovered the short-term overload risk of the room and immediately activated the emergency response mechanism. On the one hand, the system identified other non-key rooms on the same floor and implemented short-term power outages for some non-core loads such as lighting and water dispensers; on the other hand, the system calculated the power load change trend of the delivery room in real time, and controlled the temperature setting and power output of the air conditioner to ensure the safety of the delivery room while controlling the power consumption at the optimal level. This system can accurately perceive the real-time power consumption status of each room, and through hierarchical and differentiated control strategies, it can quickly intervene when local abnormalities occur, make full use of redundant power resources, and effectively avoid the occurrence of chain failures.

[0040] It can be seen that the use of the regional electricity consumption monitoring and control method in the embodiment of the present application can not only ensure the continuity of power supply for key businesses, but also effectively solve the limitations of the traditional centralized power monitoring system with coarse control granularity and lack of intelligent scheduling capabilities, thereby achieving balanced and reliable power supply.

[0041] For ease of understanding, the following describes the process of the method provided by this implementation in combination with the above scenario. Figure 1 , which is a flow chart of a method for monitoring and controlling regional electricity consumption in an embodiment of the present application.

[0042] S101. Based on the regional power network structure data and the user's power consumption settings, a multi-level power consumption model including the equipment level, the room level and the building level is constructed.

[0043] Among them, the regional power network structure data represents the topological connection relationship and equipment parameter information of the power supply and distribution network covering the target area, and refers to the structured data formed by digitally modeling the physical architecture and logical relationship of the regional power grid. User power settings refer to the parameter configurations made by users for power demand and priority in different rooms at different time periods. The equipment layer is used to represent the spatial location, rated power, power time period and other attribute information of various types of power equipment in the region. The room layer is used to represent the spatial location, floor, room type, equipment and other attribute information of each room. The building layer is used to represent the vertical height of each floor, the rooms it contains, the corresponding relationship with the distribution room and other attribute information. The multi-level power consumption model is a data framework model that can visualize the information representation and operation of the database system.

[0044] Specifically, this step is performed when the regional electricity consumption monitoring system is initially deployed or when data is updated. First, the electricity consumption monitoring system receives multi-source heterogeneous data such as CAD drawings, BIM models, GIS coordinates, etc. of the regional power network, and extracts the spatial position and topological connection relationship data of the device nodes, room nodes, and floor nodes through image recognition, semantic understanding and other technologies. Then, the extracted structured data is imported into a relational database for storage. At the same time, the system receives parameters such as the regular electricity consumption period, electricity consumption priority, and power consumption upper limit of each room input by the user through the human-computer interaction interface of a specific APP on the client, and associates them with the room nodes for storage. Finally, based on the above-mentioned regional power grid structure data and user setting data, the system generates a multi-level electricity consumption network model with three dimensions including the device layer, room layer, and building layer through methods such as tree topology modeling and relationship map construction.

[0045] In some embodiments, the construction of a multi-level electricity consumption model can be achieved in a variety of ways: optionally, first perform graphic recognition on the regional power grid drawings, extract the device locations and connection relationships therein, and then use Neo4j and other graph databases to perform multi-level network modeling; optionally, first use Autodesk Revit and other BIM software to perform semantic analysis on the building structure model, extract the room topology relationship, and then use PostgreSQL and other relational databases to organize multi-level spatial indexes. It is understandable that other modeling methods or database tools can also be used to achieve the management and application of multi-level electricity consumption models, which are not limited here.

[0046] S102: Obtain associated feature data of the target room including the number of people in the room and the equipment list, and determine the real-time business scenario of the target room based on the associated feature data.

[0047] Among them, the target room refers to the key room that needs to be monitored for abnormal power consumption, usually the key places such as the operating room and ICU in the hospital or the functional rooms with high power consumption such as the computer classroom in the school. The number of people in the room refers to the number of people in the room in real time or at a certain moment. The equipment list is used to indicate the type, model, quantity and other list information of various types of electrical equipment in the room. The real-time business scenario refers to the specific business scenario in which the target room is currently located, such as the different stages of the hospital operating room such as intraoperative, between-operation, and vacant. The associated feature data represents the data set after associating the number of people in the room and the equipment list, which comprehensively describes the power consumption characteristics of the room. For example, the associated feature data can be {"roomID":"ConferenceRoom101","timestamp":"2023-05-20T14:30:00Z","occupancy":12,"deviceList":[{"type":"Projector","model":"A","power":500,"count":1,"status":"on"},{"type":"Computer","model":"B","power":200} , "count": 3, "status": "on"}, {"type": "Light", "model": "C", "power": 100, "count": 10, "status": "on"}, {"type": "AirConditioner", "model": "D", "power": 1500, "count": 1, "status": "on"}]}, The associated feature data in JSON format includes the target conference room ID, data collection timestamp, number of people in the room, and a detailed equipment list. The equipment list further lists the attributes of each type of equipment, such as model, single-machine power, quantity, and operating status.

[0048] Specifically, the electricity monitoring system collects image data or thermal imaging data of people entering and leaving the target room through video cameras or infrared sensors installed at the door and corridor of the target room, and uses target detection and tracking algorithms to count the number of people in the room in real time. At the same time, the system connects with business systems such as the hospital HIS (Hospital Information System) and EMR (Electronic Medical Record) to obtain information such as the equipment category, model, power parameters, etc. of the target room, and form a structured room equipment list. In addition, the system also maps the department to which the room belongs and its corresponding business process through the knowledge graph, and combines data such as the equipment list, number of people, and appointment information, and uses a scene recognition model based on rules or machine learning to determine the most likely business scenario in which the room is currently located.

[0049] In some embodiments, the associated characteristic data may also include associated environmental data in the room. That is, various sensors connected to the Internet of Things may be used to obtain the temperature, humidity, brightness, and door and window status in the room; the current weather status or holiday period and time may also be obtained through a specific weather API interface on the network, so as to make a more accurate judgment on the real-time business scenario of the target room.

[0050] In some embodiments, the associated features of the room can be obtained and scene recognition can be achieved in a variety of ways: Optionally, use OpenPose and other frameworks to estimate human postures in surveillance video frames, and determine the number and activity status of people in the room through key point coordinates; combine the identity tags of doctors, nurses, and patients to build a medical business process finite state machine, and use HMM and other models to achieve scene classification. Optionally, use YOLOv3 and other algorithms to perform target detection on medical device images in the room to obtain the location and quantity information of various types of equipment; associate and map the detection results with the business attributes of the department, and use models such as random forests for scene reasoning. It is understandable that other multimodal perception, feature fusion and scene understanding technologies can also be used to achieve business scene recognition in the target room, which is not limited here.

[0051] S103: Based on the historical electricity consumption data of multiple similar rooms in the room layer on the multi-level electricity consumption model, determine the electricity consumption data boundary of the target room in the real-time business scenario.

[0052] Among them, multiple similar rooms in the room layer of the multi-level power consumption model refer to other rooms in the room hierarchy of the power consumption network model that belong to the same floor as the target room and have similar room functional attributes. Historical power consumption data represents the time series record of power consumption parameters such as power and current over a period of time in the past. The power consumption data boundary is used to represent the power consumption parameter threshold range corresponding to the normal power consumption behavior of the target room in a certain business scenario.

[0053] Specifically, the electricity consumption monitoring system first retrieves other rooms of the same floor and room type as the target room from the room dimension of the multi-level electricity consumption model. Then, the system queries the historical power consumption, current and other time series data of these similar rooms in the same business scenario. Next, the system performs statistical analysis on the historical data of these similar rooms, and uses percentile, standard deviation and other algorithms to remove outliers to obtain the aggregated feature values ​​in this scenario. Finally, the system shifts the upper and lower limits of the aggregated feature values ​​by a certain proportion to form the upper and lower boundaries of the target room's electricity consumption data in the current real-time scenario. This boundary reflects the normal electricity consumption range of most rooms under the same conditions, and can be used to evaluate whether the target room's electricity consumption behavior is abnormal.

[0054] In some embodiments, the electricity usage boundary of a room can be determined in a variety of ways: Optionally, the historical data of a certain electricity usage indicator of similar rooms in the target scenario is clustered, the center point of each cluster is extracted as multiple candidate thresholds, and then a voting mechanism is used to determine the threshold with the highest frequency of occurrence as the boundary value. Optionally, a deep learning model such as LSTM is used to fit the electricity usage curve of similar rooms in this scenario, and the peak and valley values ​​of the fitted curve are used as the upper and lower limits of the boundary respectively. It is understandable that other algorithms such as time series data mining and boundary learning can also be used to realize the extraction of the electricity usage behavior portrait of the target room, which is not limited here.

[0055] S104. When it is detected that the real-time power consumption data corresponding to the target room exceeds the power consumption data boundary, the target room is regarded as an abnormal room, and abnormal equipment and abnormal floors corresponding to the abnormal room are determined based on a multi-level power consumption model.

[0056] Among them, real-time power consumption data refers to the measured values ​​of power consumption, current and other parameters of the target room at the current moment. Abnormal room refers to the target room whose real-time power consumption behavior deviates from the normal mode. Abnormal equipment is used to indicate the faulty or overloaded electrical equipment that causes abnormal power consumption in the room. Abnormal floor refers to the floor area with one or more abnormal rooms.

[0057] Specifically, the power consumption monitoring system first compares the real-time parameters with the threshold boundaries. If any parameter exceeds the limit, the target room is marked as an abnormal room. Then, the system analyzes the real-time status data of the device nodes connected to the abnormal room in the multi-level power consumption model, identifies the devices that are likely to cause abnormalities through rule reasoning or correlation analysis, and marks them as suspected abnormal devices. Next, the system searches for the floor to which the abnormal room belongs in the floor dimension of the model. If the floor contains other abnormal rooms, the floor is marked as an abnormal floor. Finally, the identification results of abnormal devices, abnormal rooms, and abnormal floors are output and alarms are issued.

[0058] In some embodiments, the determination of abnormal equipment and abnormal floors can be achieved in a variety of ways: Optionally, abnormal equipment can be determined directly from the end-to-end power consumption data of abnormal rooms using characteristic rules such as current and voltage mutations and harmonic distortion; the floor area nodes corresponding to the abnormal rooms in the topological network are inferred using a graph model, and the floor abnormality is determined based on the node status. Optionally, an association rule mining algorithm is used to calculate the correlation between the power consumption parameters of the abnormal room and the equipment therein, and the equipment with the highest confidence is determined to be abnormal; the severity of the abnormality of each room on the floor is weighted and scored, and if the total score threshold is exceeded, it is determined to be an abnormal floor. It is understandable that other data mining and fault diagnosis technologies can also be used to achieve the function of tracing back the device-level root cause and floor-level impact from the abnormal power consumption in the room, which is not limited here.

[0059] S105. After determining abnormal features of the abnormal device according to the real-time business scenario and the real-time power consumption data, generate a device optimization instruction based on the abnormal features and send the device optimization instruction to the abnormal device.

[0060] Among them, abnormal characteristics refer to the key parameters such as power consumption, current waveform, harmonic content, etc. of abnormal equipment when an abnormality occurs, which are significantly different from the normal state. Equipment optimization instructions refer to control instructions for adjusting the operating parameters or control logic of the equipment in order to suppress abnormal characteristics and restore normal operation of the equipment.

[0061] Specifically, the power consumption monitoring system first matches the optimization measures applicable to the device from the preset device optimization strategy library based on the type, model and other attributes of the abnormal device. Then, the parameter adjustment range, logical switching conditions and other contents involved in the measures are filled into the predetermined instruction template to generate executable instructions that can be directly issued to the device control unit. If there is no match in the strategy library, the system can also call the alarm handling solution provided by the device to generate instructions. Finally, the instructions are sent to the abnormal device via a wired or wireless network, and it responds. For devices that have not recovered after the execution of the instructions, the system will issue a warning and recommend manual handling.

[0062] In some embodiments, the generation and issuance of equipment optimization instructions can be achieved in a variety of ways: Optionally, the power monitoring system uses an expert knowledge base to determine treatment measures for common anomalies of various types of equipment, and then uses a template matching algorithm to generate parameterized instructions executable by the equipment in real time; the instructions are issued to field equipment through industrial protocols such as Modbus. Optionally, data-driven methods such as deep reinforcement learning are used to make real-time decisions on optimal treatment measures based on equipment anomaly characteristics and historical adjustment experience; the device's own remote control interface is called to convert the measures into native device instruction codes and issue them. It is understandable that other rule engines, policy optimization, and instruction adaptation technologies can also be used to achieve a closed-loop mapping of abnormal diagnosis results to executable actions for equipment, which is not limited here.

[0063] S106. Based on the abnormal characteristics and the proportion of power consumption of the abnormal rooms on the abnormal floor, control the power supply scheduling of multiple room layers on the abnormal floor.

[0064] Among them, the proportion of power consumption of abnormal rooms in abnormal floors represents the percentage weight of abnormal room power, current and other parameters in the summary value of the entire floor. Power supply scheduling refers to the dynamic optimization control of redistributing limited power resources by adjusting the power distribution topology and power supply priority.

[0065] Specifically, the power consumption monitoring system first evaluates the proportion of end-to-end power consumption characteristics of abnormal rooms in the floor dimension of the multi-level power consumption model, and divides them into key abnormal rooms or general abnormal rooms accordingly. Then, the system formulates emergency plans such as regional rotation and peak power consumption based on the business and power consumption correlation of each room on the floor. Next, the plan is refined into specific operating steps such as switching switches and adjusting loads to form a policy configuration file that can be read by the building control system. Finally, the configuration file is transmitted to the building control system, which directly controls the switches in the cabinet to complete the jump, or parses the strategy into control instructions and sends them to the end load equipment for execution. During the entire scheduling process, power supply continuity is always ensured for key medical protection areas.

[0066] In some embodiments, floor power supply dispatching control can be achieved in a variety of ways: Optionally, based on the power consumption and business importance attributes of abnormal rooms, a fuzzy comprehensive evaluation model is used to determine its abnormal level; a hierarchical control plan is formulated, and measures such as timed rotation and service degradation are taken, and the building control system is configured to execute through the BACnet protocol. Optionally, a robust optimization algorithm is used to solve a combination strategy that considers multi-room power consumption constraints and scheduling costs; the solution results are parsed into equipment operation instructions that can be directly driven by the building control, and issued through broadcast or multi-point telemetering. It is understandable that other load classification, emergency strategy optimization and building automation technologies can also be used to achieve rapid isolation of local abnormalities on the floor and overall coordinated control, which is not limited here.

[0067] In the above embodiments, the optimization control of device-level abnormalities is mainly achieved by adjusting device operating parameters, updating control logic, or cutting off power supply. In practical applications, the system can also deeply explore the correlation characteristics between the professional attributes of medical equipment and the environment and personnel behavior, form a domain knowledge base, and conduct pre-emptive prevention through situational awareness and behavior guidance before the device abnormality occurs.

[0068] The following is a supplement to the scenario of this embodiment.

[0069] During the emergency response process in the obstetrics and gynecology delivery room, the power consumption monitoring system further explored the correlation characteristics between the room environment data and the operating parameters of medical equipment. For example, when the room temperature is higher than 26 degrees Celsius, the acquisition accuracy of the fetal heart monitor will drop by 2 percentage points; when the air humidity is lower than 40%, the heating power of the neonatal incubator will increase by 15%. Based on these correlation characteristics, the system dynamically adjusts the air supply temperature and humidity of the air conditioner to minimize unnecessary energy consumption while meeting the comfort requirements, and by optimizing the working mode of the incubator and the aspirator, it minimizes the instantaneous power impact while ensuring performance, further improving the reliability of key equipment.

[0070] In addition, the system also pushes real-time alarm information about the room environment and equipment status to medical staff, recommends reasonable adjustment of equipment usage time, and guides newborns' families to visit at different times to stabilize the room's power load to the greatest extent. Based on intelligent perception and collaborative optimization, this system can also use medical scenario data to discover the operation rules of equipment and form a professional field knowledge base.

[0071] After combining the above scenarios, the following is a more detailed description of the process of the method provided by this implementation. Figure 2 , is another flow chart of the method for monitoring and controlling regional electricity consumption in an embodiment of the present application.

[0072] S201. Based on the regional power network structure data and the user's power consumption settings, a multi-level power consumption model including the equipment level, the room level and the building level is constructed.

[0073] Referring to step S101 , the power consumption monitoring system constructs a multi-level power consumption model.

[0074] In some embodiments, the electricity consumption monitoring system will: receive and parse regional power network structure data, determine the device nodes, room nodes, floor nodes and topological connection relationships between the nodes in the regional power network structure data; access the electricity consumption information database to extract the attribute information of each node; the attribute information includes the rated parameters of the equipment, the area of ​​the room, the functional attributes and the vertical position information of the floor; generate a multi-level electricity consumption topology network based on the topological connection relationship and the attribute information; the multi-level electricity consumption topology network is a hierarchical structure including the device layer, the room layer and the floor layer; receive the electricity consumption settings entered by the user in the user terminal, and use the electricity consumption settings as constraints to map and label the multi-level electricity consumption topology network; the electricity consumption settings include the electricity consumption time period of each room, the rated power of each device and the electricity consumption priority parameters; based on the labeled multi-level electricity consumption topology network, construct a multi-level electricity consumption model.

[0075] Specifically, the regional power network structure data represents the network topology data with power distribution equipment as nodes and line connections as edges. The power consumption information database refers to a data set that stores the static attributes of power consumption objects such as rooms and floors. The multi-level power consumption topology network refers to a networked data structure that contains power consumption objects of multiple granularities such as equipment, rooms, and floors, and reflects their physical and logical associations. Power consumption settings represent the user's configuration constraints on the attributes and behaviors of various types of power consumption objects.

[0076] The electricity consumption monitoring system will first extract the device node information from the topological structure data of the regional power grid, and build a multi-level mesh data structure by mapping the location with the room and floor. Then, the static attribute labels of the equipment, room, and floor are supplemented from the electricity consumption information library, and these attributes are attached to the corresponding network nodes. Next, the constraints such as power consumption period, power, priority, etc. set by the user are pressed into the multi-level network, and the attributes of each node and connection are annotated and transferred. Finally, using the annotated multi-level topological network as a sample, through machine learning methods such as representation learning, an end-to-end multi-level electricity consumption model that can reflect the association pattern in the electricity consumption data is constructed.

[0077] Take the power consumption system of an office building as an example. First, the power consumption monitoring system extracts the connection topology data of each distribution room, switch cabinet, and power supply line of the building from the distribution automation system and maps it into a device layer topology network. Then, each power supply circuit is associated with the room according to the floor plan, and the room layer topology is abstracted. Then, the rooms are aggregated to their respective floors according to the vertical distribution of the floors to form a building layer topology. In this way, a multi-level network of "equipment-room-floor" is constructed. Next, the power consumption monitoring system accesses the property management system to obtain the attributes such as the area, functional division, and building height of each floor of each room, and injects the corresponding network nodes. Then, the power consumption monitoring system receives the constraint rules set by the property management personnel on the energy consumption management terminal, such as the time interval for each room to use electricity, the meter capacity of each circuit, and the priority guarantee of electricity use in important office areas, and embeds them into the multi-layer network. Finally, the power consumption monitoring system uses a graph neural network to embed the nodes and edges of the multi-layer network into the same low-dimensional space, and extracts the unified features of power consumption behaviors of different granularities through end-to-end representation learning. On this basis, a multi-level electricity consumption model is constructed through data mining techniques such as feature clustering, association analysis and trend extrapolation.

[0078] S202: Obtain associated feature data of the target room including the number of people in the room and the equipment list, and determine a real-time business scenario of the target room based on the associated feature data.

[0079] Referring to step S102 , the electricity usage monitoring system determines the real-time business scenario.

[0080] In some embodiments, the electricity consumption monitoring system will: obtain a real-time image of the target room, and determine the number of people in the room based on the real-time image; determine the equipment list of the target room in combination with the real-time image and user settings; the equipment list includes the equipment model information, power parameters and quantity information of various electrical equipment configured in the target room; use the number of people in the room and the equipment list as associated feature data, and determine the business adaptation set according to the target room and the associated feature data; input the associated feature data into the business scenario recognition model corresponding to the business adaptation set to obtain the real-time business scenario; when the number of people in the room or the equipment list changes, re-acquire the changed associated feature data, and dynamically update the real-time business scenario of the target room.

[0081] Specifically, real-time images refer to real-time video frames inside a room collected by visual sensors such as cameras. The number of people in a room refers to the total number of human targets detected in the current room environment. The equipment list indicates the type, specifications, power consumption and other attribute information of the electrical equipment actually configured and used in the room. The service adaptation set indicates the category of alternative service scenarios matched according to the room attributes.

[0082] The business scenario recognition model refers to a machine learning classification model that takes associated features as input and business scenarios as output. This model is used to determine which business scenario the room is most likely to be in based on the real-time status of the people and equipment in the room. The training and use process of the model is as follows: First, the system collects real-time video images through the cameras deployed in the room, and uses the target detection and tracking algorithm to count the number of people in the room. At the same time, the business system is accessed to obtain the type, quantity, power and other parameters of the electrical equipment in the room, and a real-time equipment list is generated. Then, the system matches several possible business scenarios from the knowledge base as a candidate set based on fixed attributes such as the area, number of people accommodated, and floor to which it belongs. Next, the current real-time features such as the number of people and equipment status are input into a multi-classifier integrated model to score the confidence of each candidate scenario. The classifier uses multiple machine learning algorithms such as KNN, naive Bayes, and decision trees, and is trained with historical data as samples. The input features are the real-time number of people and equipment combinations in the room, and the output is the most likely business scenario category. Finally, the scene with the highest confidence is taken as the discrimination result. Considering that the room status is dynamically changing, when the number of people or equipment changes, the system will re-trigger the above process and perform incremental learning on the original classifier to continuously adapt to the new data distribution. The specific training process is: collect a large amount of historical data such as the number of people and equipment configuration in different business scenarios in the room, and label each data sample with a scene. Then divide the training set and the test set, use the training set samples to fit and learn the model parameters of each classifier, and then use the test set to evaluate the model performance. The hyperparameters of the model are optimized through grid search and other methods to improve the generalization ability. The evaluation indicators used include precision, recall rate, F1 value, etc. The input of the model is the real-time status characteristics of the room, including population statistics and equipment lists. The output is the confidence score for each alternative business scenario and the scene category with the highest confidence. Through this model, the system can perceive the business semantics of the room in real time and provide context information for subsequent power consumption anomaly analysis and scheduling optimization.

[0083] The electricity monitoring system will collect room video image data in real time by deploying cameras in the room. Then, it will use the target detection and tracking algorithm to continuously count the changes in the number of people in the room. Next, the electricity monitoring system will analyze the usage status of the electrical equipment in the image and generate a real-time equipment usage list and energy consumption profile of the room in combination with the user configuration. Then, the number of people in the room and the equipment list are used as contextual features, and a set of alternative business scenarios are determined by associating and matching with the room's own attributes. The electricity monitoring system will then use the associated features as input, and through pattern recognition technologies such as vector nearest neighbors and Bayesian fusion, it will identify the most likely business scenario category of the room from the alternative set. Considering that the status of people and equipment in the room changes dynamically, the system will also monitor the changes in associated features in real time, triggering real-time retraining of the scene recognition model, thereby dynamically updating the semantic label of the room.

[0084] Take a conference room as an example. The power consumption monitoring system first collects the internal images of the conference room through the indoor camera in real time, and uses real-time target detection frameworks such as YOLOv3 to identify human targets in the image. Through inter-frame tracking of the target detection results, the system counts the total number of people in the conference room in real time. Then, the system uses an electrical appliance recognition model based on image classification to detect the power-on status of devices such as projectors and computers in the image. Then, the conference room attribute library is accessed to obtain the rated power of the electrical equipment configured in the conference room. This forms a real-time equipment list for the conference room and estimates its total power consumption requirements. Then, based on the static attributes of the conference room such as the area, the number of people accommodated, and the floor on which it is located, the system matches possible business scenario sets such as "department regular meeting, lunch misappropriation, and idleness" in the business rule library. Then, using the current real-time features such as the number of people and device status as input, the system uses multiple classifiers such as KNN and Bayes to identify the real-time business scenario with the highest confidence from the candidate set, such as "department regular meeting". It is worth noting that if people enter and exit or the status of the equipment changes during the meeting, resulting in changes in the number of people and equipment list features, the system will re-trigger the scene recognition process. First, the associated feature vector is updated and input into the trained classification model for re-prediction. At the same time, the forgetting mechanism is introduced to fine-tune the original model and continuously learn new business models. This achieves real-time perception and adaptive optimization of conference room business scenarios.

[0085] In some embodiments, the electricity consumption monitoring system will: preset multiple corresponding adaptation business scenarios according to the room type and room purpose of the target room; in the adaptation business scenarios, filter out a first business subset with a number of people higher than a first threshold according to the number of people in the room; obtain the historical equipment list of the room, compare the historical equipment list with the equipment list, and obtain a new equipment list and a reduced equipment list; if the new equipment list and the reduced equipment list are both empty, obtain the historical business scenarios of the target room at the previous moment, and filter out a second business subset whose deviation from the historical business scenarios is less than a second threshold from the first business subset according to the historical business scenarios; if the new equipment list is not empty, calculate the matching degree between the equipment in the new equipment list and the standard equipment corresponding to each business scenario in the first business subset, and filter out a second business subset with a matching degree greater than a third threshold from the first business subset according to the matching degree; if the reduced equipment list is not empty, remove the business scenarios that include the equipment in the reduced equipment list from the standard equipment corresponding to each business scenario in the first business subset to obtain a second business subset; and determine the second business subset as the business adaptation set of the target room.

[0086] Specifically, the historical equipment list records the electrical equipment used in the room over a period of time. The newly added equipment list and the reduced equipment list respectively indicate the electrical appliances that have appeared and disappeared in this statistics compared with the history. The historical business scenario refers to the business scenario type identified in the room at the last statistical moment. The deviation quantifies the degree of semantic difference between two business scenarios. Standard equipment refers to the set of typical electrical equipment that is usually equipped for each business scenario. The business adaptation set refers to the set of adaptation candidate scenarios obtained through multiple rounds of screening.

[0087] The matching degree is the similarity between the equipment list actually configured in the target room and the standard equipment list corresponding to a candidate business scenario. The specific method for calculating the matching degree includes: extracting feature vectors from the real-time equipment list of the room and the standard equipment list of the candidate business scenario, and then calculating the similarity of the two feature vectors. The feature vector of the equipment list usually contains information such as the type, model, and quantity of various types of equipment. The similarity can be measured by cosine similarity, Euclidean distance, etc. in the vector space model.

[0088] The first threshold refers to the minimum number of people required when matching business scenarios. This threshold is obtained by statistically analyzing massive historical business data to obtain the distribution characteristics of the number of people corresponding to different business scenarios, and then taking a specific quantile (such as the median) as the first threshold. If it is higher than the first threshold, it means that the number of people in the room meets the basic requirements of the scenario.

[0089] The second threshold refers to the maximum deviation of a business scenario that is considered similar to a historical scenario. The system uses a specific similarity metric to calculate the degree of difference between the current business scenario and the historical scenario. If it is less than the second threshold, the scenario is considered to have not changed substantially. The second threshold is determined by clustering analysis of historical data, and the scenarios in the same cluster are considered to be of the same type. The maximum similarity between the scenarios in the same class is calculated as the second threshold.

[0090] The third threshold refers to the minimum matching degree requirement between the actual equipment in the room and the standard equipment of the candidate scenarios. The system compares the matching degree between the newly added equipment and each candidate scenario, and the scenarios above the third threshold are screened. The third threshold is obtained by analyzing historical business data, counting the distribution of equipment matching degrees of the correct business scenarios, and taking a specific quantile as the minimum acceptable standard.

[0091] For example, the equipment list of a conference room is represented as a multidimensional vector. Suppose we consider four types of equipment: projectors, computers, lighting, and air conditioners, and their corresponding rated powers are 500W, 200W, 100W, and 1500W, respectively. The actual equipment list of conference room A at the current moment is: projector: 1, computer: 3, lighting: 10, air conditioner: 1; then the actual equipment vector of conference room A is represented as:

[0092] X=(1,3,10,1)×(500,200,100,1500)=(500,600,1000,1500).

[0093] Let's assume that we have obtained the standard equipment list for a "large conference" scenario through statistics of historical data: 2 projectors, 5 computers, 15 lightings, and 2 air conditioners; then the corresponding standard equipment vector is:

[0094] Y=(2, 5, 15, 2)×(500, 200, 100, 1500)=(1000, 1000, 1500, 3000), then the cosine similarity between the actual device vector X of conference room A and the standard device vector Y of "large conference" is: cosθ==500*1000+600*1000+1000*1500+1500*3000 / (sqrt(500*500+600*600+1000*1000+1500*1500)*sqrt(1000*1000+1000*1000+1500*1500+3000*3000))≈0.993;

[0095] Suppose that we have determined the third threshold to be 0.98 through historical data analysis, that is, the match between the actual equipment list and the standard list must reach 0.98 or above to consider that the room is in the business scenario corresponding to the standard; and the similarity in this scenario is 0.993, which meets the requirements and is judged as a "large conference" scenario.

[0096] The first, second and third thresholds mentioned above may also be set and modified by the developer by default, or modified by the user.

[0097] The power consumption monitoring system will first match a set of standard business scenarios with high adaptability in the knowledge base as a candidate set based on the inherent attributes of the room, such as area, floor, partition, etc. Then, the system will continuously reduce the size of the adaptation set through two rounds of dynamic screening to obtain the business scenario that best matches the real-time status of the room. In the first round of screening, the system first calculates the matching degree of each candidate scenario in terms of the number of people based on the real-time room population statistics, and screens out the scenarios with too large a difference in the number of people. In the second round of screening, the system first obtains the historical equipment usage records of the room for a period of time before, and compares them with the real-time equipment list of this time to obtain the changes in the newly added and reduced equipment. If the equipment has not changed, the system tends to believe that the business scenario of the room has not changed much compared with the previous moment, so the historical scenario of the previous moment is used as the benchmark to screen out a subset of scenarios with a smaller deviation from it. If a new device appears, the system calculates the matching degree of the new device with the standard equipment of each candidate scenario, and screens out the scenarios with a higher degree of matching. If a reduced device appears, the system directly eliminates those business scenarios that must use the device. Finally, the results of the two rounds of screening are used as the real-time business adaptation set of the target room.

[0098] S203: Based on the historical electricity consumption data of multiple similar rooms in the room layer on the multi-level electricity consumption model, determine the electricity consumption data boundary of the target room in the real-time business scenario.

[0099] Referring to step S103 , the power consumption monitoring system determines the power consumption data boundary.

[0100] S204: When it is detected that the real-time power consumption data corresponding to the target room exceeds the power consumption data boundary, the target room is regarded as an abnormal room, and abnormal equipment and abnormal floors corresponding to the abnormal room are determined based on a multi-level power consumption model.

[0101] Referring to step S104, the electricity monitoring system will determine the abnormal equipment and abnormal floor corresponding to the abnormal room.

[0102] S205. Determine various power usage parameters of abnormal equipment in multiple power usage time periods based on real-time business scenarios and real-time power usage data.

[0103] Among them, the power consumption period refers to the division of 24 hours a day into several discrete time intervals according to the power load characteristics, such as peak, flat and valley, etc. Power consumption parameters refer to the key physical quantities that describe the power consumption behavior of equipment in a certain period of time, and the common ones are useful power, current, voltage and power factor, etc.

[0104] Specifically, the power consumption monitoring system first extracts the historical power consumption curves of abnormal equipment in different time periods from the multi-level power consumption model, and compares them with the standard curves of similar equipment in the same business scenario time by time. Then, the system uses digital signal processing technology to extract features from the real-time power consumption data stream of abnormal equipment to obtain multi-dimensional parameters such as average power, peak current, and tracking coefficient in each time period. Next, the system calculates the degree of deviation of these parameters from the standard values ​​and generates a time period parameter matrix that reflects the performance of the equipment. Finally, through clustering or correlation analysis of the matrix, the power consumption monitoring system can identify the key time periods and power consumption parameters that have the most prominent impact on the abnormal status of the equipment.

[0105] In some embodiments, the time period measurement of abnormal power consumption parameters of equipment can be achieved in a variety of ways: Optionally, by integrating with a power collection terminal or smart gateway, the multi-channel power consumption and environmental measurement data of the equipment are collected at a minute-level granularity; the collected data is decomposed into time series, and the contribution of different frequency components in each time period is extracted as characteristic parameters. Optionally, a convolutional neural network is used to model the real-time power consumption curve of the equipment, and its local morphological characteristics are characterized by model parameters; a gated recurrent unit network is used to perform multi-step prediction of the power consumption curve, and its global trend characteristics are characterized by prediction errors. It is understandable that other digital twins, data dimensionality reduction and feature engineering technologies can also be used to achieve a refined power consumption behavior portrait of abnormal equipment, which is not limited here.

[0106] S206. Determine the abnormal characteristics of the abnormal equipment including equipment performance degradation, power risk level and energy efficiency level based on various power usage parameters.

[0107] The degree of equipment performance degradation indicates the percentage by which the actual working capacity of the equipment is reduced compared to the rated level. The power risk level refers to the possibility of a safety accident caused by abnormal performance of the equipment. The energy efficiency level is used to indicate the energy consumption intensity per unit output of the equipment, usually measured by electricity consumption or carbon emission equivalent.

[0108] Specifically, the electricity monitoring system first compares and analyzes the percentage difference between the power and current parameters in each time period and the rated value, estimates the performance degradation of the equipment under different load conditions, and obtains the average degradation index. Then, the system uses the preset equipment risk assessment rule library to calculate the degree of proximity of the parameters in each time period to the risk threshold one by one, and obtains the overall electricity risk level of the equipment by weighted average. Next, the system combines the rated energy consumption and actual output of the equipment, calculates its unit energy efficiency value in different time periods, and then maps it to the energy efficiency identification grade system to determine its grade. Finally, the degradation degree, risk level, and energy efficiency level are output as the three core characteristic dimensions of abnormal equipment operation.

[0109] In some embodiments, the conversion from power consumption parameters to abnormal features can be achieved in a variety of ways: Optionally, a random forest algorithm is used to establish an abnormal feature analysis model for time period parameters and performance degradation, and the most relevant parameters are screened using feature importance scores.

[0110] The abnormal feature analysis model is used to extract key features that reflect performance degradation and energy efficiency decline from the abnormal power consumption data of the equipment. The training and use process of the model is as follows: First, the system obtains the monitoring data of power, current, voltage, etc. of the equipment in each power consumption period, as well as the standard reference values ​​of similar equipment under the same working conditions. Then, digital signal processing technology is used to reduce noise, compress and extract features of the measured data to obtain the statistical values ​​of various parameters of the equipment in different periods, such as average power, peak-to-valley ratio, etc. Next, the degree of deviation of these statistical values ​​relative to the standard value is calculated to generate multi-dimensional time series features reflecting the performance of the equipment. On this basis, the system uses machine learning algorithms to establish a mapping relationship between characteristic parameters and equipment health status. Taking the historical working condition data of a large number of similar equipment as training samples, each piece of data contains the parameter characteristics of a certain equipment at a certain time and the corresponding health status label, such as "performance degradation degree = 15%, energy efficiency level = 2, power consumption risk = high", etc. The system uses these data to train a regression model or classification model so that it can evaluate the quantitative level of multiple abnormal indicators based on the real-time characteristics of the equipment. Commonly used algorithms include random forest, support vector machine, neural network, etc.

[0111] Specifically, the training method of the abnormal feature analysis model is as follows: historical samples are grouped by device type, and each group is further divided by time series. For the time series data of each device, the statistical features within several windows are extracted by sliding. The health status of the data in each time window is labeled by an expert or rule engine. Features and labels are combined into training sample pairs, and the training set and test set are split using methods such as cross-validation. The model parameters are fitted with the training set, and the performance is evaluated with the test set. Loss functions such as mean square error and cross entropy, and evaluation indicators such as accuracy and area under the ROC curve are used. Features and models are continuously debugged, and after the optimization indicators are met, the model parameters are saved for online calling. The input of the model is the real-time multi-parameter time series data of the device, and the output is the abnormal feature indicators such as the performance degradation degree, energy efficiency level, and risk level of the device. These indicators can intuitively present the health status of the equipment and provide a quantitative basis for further abnormal cause diagnosis and prescription decision-making.

[0112] S207: According to the abnormal characteristics, a device optimization strategy is matched in a preset strategy library, and a corresponding device optimization instruction is generated based on the device optimization strategy.

[0113] The preset strategy library represents a collection of equipment abnormality handling solutions predefined based on expert experience. Equipment optimization strategy refers to a set of actions to correct abnormal behavior and improve operation quality by adjusting the controllable parameters of the equipment and updating the control logic of the equipment.

[0114] Specifically, the power consumption monitoring system first introduces the feature vector into the matching engine of the preset strategy library, and finds several strategies that best fit the current abnormal scenario through similarity calculation. Then, the system simulates and deduces these candidate strategies and pre-evaluates the effects, and selects the strategy with the best benefit-risk ratio based on the device's own constraints. Next, the system selects the corresponding control instructions from the instruction template library based on the strategy content, and fills in the actual device parameter thresholds, logical condition judgment statements, etc., to generate a device-level optimized instruction set that can be directly executed. Finally, the instruction set is transmitted to the field controller of the device layer through the standard industrial bus protocol. The controller then parses the instruction semantics, plans the execution steps, and calls the underlying driver to realize the conversion of instructions to device actions.

[0115] The device optimization strategy matching model can be used to match the device optimization strategy. The model is used to quickly search the knowledge base and find the best abnormal handling measures according to the abnormal characteristics of the device.

[0116] The training and use process of the equipment optimization strategy matching model is as follows: First, the system builds an abnormality handling strategy knowledge base offline, collects the experience of equipment experts and operation and maintenance personnel, maps each common equipment abnormality problem with its corresponding handling measures one by one, and forms a paired strategy rule. The handling measures can be to adjust the operating parameters, replace parts, limit the load, etc. Each rule indicates the applicable equipment type, abnormal mode, effect evaluation and other attribute labels. When applied online, for a newly discovered equipment abnormality, the system first extracts its abnormal feature vector, including failure mode, performance index deviation, etc. Then, use the feature vector as the query (query parameter) to perform a fast matching search in the knowledge base. To improve retrieval efficiency, the system pre-indexes the strategy library and trains a variety of similarity models. Common similarity measurement methods include cosine similarity, Euclidean distance, Jaccard coefficient, etc. For example, word embedding models such as word2vec can convert abnormal feature semantics into dense vectors for fast matching.

[0117] The specific matching process is as follows: perform a one-to-many similarity calculation between the device abnormality feature vector and the feature label of each strategy in the knowledge base, and select multiple most relevant strategies. Then perform fine-grained matching and filtering on the candidate strategies, and evaluate the actual effect and feasibility of each strategy based on factors such as the physical state of the device, environmental constraints, and cost budget, and select the best strategy. If necessary, start simulation deduction to predict the results of strategy execution. Finally, convert the strategy into executable control instructions and send them to the device.

[0118] The input of the model is the abnormal feature vector of the equipment and the environmental constraints, and the output is the optimal disposal strategy for specific abnormal situations. By mining the strategy knowledge base obtained from expert experience offline and combining it with the online fast matching algorithm, the system can dynamically recommend diagnostic prescriptions based on the actual abnormal state of the equipment and realize abnormal closed-loop processing.

[0119] In some embodiments, the generation of abnormal features to optimization instructions can be achieved in a variety of ways: optionally, using a rule-based reasoning engine to achieve strategy matching, and using a knowledge graph to represent the semantic connection between the abnormal state of the equipment and the disposal measures; using template-based code generation technology to achieve automatic conversion from strategy to instructions, and defining the equipment control logic through a domain-specific language. Optionally, the state-action value function in reinforcement learning is used to fit the mapping between the abnormal characteristics of the equipment and the optimal control strategy; natural language processing technology is used to extract key action semantics from the strategy text, and then assembled into serialized instructions that can be executed by the equipment. It is understandable that other knowledge engineering, intelligent planning and industrial software technologies can also be used to achieve a closed loop of equipment operation optimization based on abnormal diagnosis results, which is not limited here.

[0120] In some embodiments, the electricity monitoring system will: obtain a first optimization strategy for device layer abnormalities, a second optimization strategy for room layer abnormalities, a third optimization strategy for floor layer abnormalities, a fourth optimization strategy for building layer abnormalities, and a fifth optimization strategy for area layer abnormalities according to abnormal feature matching in a preset strategy library; the first optimization strategy is used to indicate adjustment of operating parameters of abnormal equipment or temporary cutting off of its power supply; the second optimization strategy is used to indicate power limitation or timed power outage for abnormal equipment in the room layer; the third optimization strategy is used to indicate total control or time-sharing power supply for abnormal rooms in the building layer; the fourth optimization strategy is used to indicate power load balancing between floors in the building layer; the fifth optimization strategy is used to indicate power scheduling for buildings in the area layer; according to the abnormal level corresponding to the abnormal feature, determine the target optimization strategy among the first optimization strategy, the second optimization strategy, the third optimization strategy, the fourth optimization strategy, and the fifth optimization strategy; according to the target optimization strategy, generate equipment optimization instructions for the corresponding level, and send the equipment optimization instructions to the node devices of the corresponding level.

[0121] Specifically, the preset strategy library refers to a series of standard processing strategies set in advance according to different granularities and types of power consumption anomalies. Abnormal features refer to abnormal power consumption patterns detected by analyzing power consumption curves. The equipment layer, room layer, floor layer, building layer, and regional layer represent different monitoring levels of the power consumption monitoring system divided according to spatial granularity. The optimization strategy includes processing measures for anomalies at different monitoring levels, such as adjusting operating parameters, power limiting, timed power off, load balancing, and power scheduling. The abnormal level refers to the minimum spatial granularity at which anomalies are detected. The target optimization strategy refers to the best processing measure matched according to the abnormal level. The equipment optimization instruction represents the control instruction required to implement the optimization strategy.

[0122] The power consumption monitoring system presets matching standard optimization strategies for power consumption anomalies of different monitoring granularities to form a strategy library. When the system detects a certain abnormal feature in the power consumption curve, it first determines the spatial granularity of the anomaly, that is, the abnormal level. Then, it searches for the treatment measures for the abnormal level in the preset strategy library. The lower the abnormal level, the finer the granularity of the optimization strategy and the more specific the treatment measures. For abnormalities at the device level, the optimization strategy tends to adjust the operating parameters of a single device or directly cut off its power supply; for abnormalities at the room level, the strategy tends to limit the power of the electrical equipment in the room or cut off the power at a scheduled time; for abnormalities at the building level, the strategy tends to control the total power consumption of the floor or supply power in different time periods; for abnormalities at the building level, the strategy tends to balance the power load between floors; for abnormalities at the regional level, the strategy tends to dispatch power to buildings in the entire area. The system automatically matches the treatment strategy of the corresponding granularity according to the actual detected abnormal level to form the best target optimization strategy combination. Finally, the optimization strategy is converted into executable device control instructions and sent to node devices at the corresponding level, such as smart meters, power supply lines, distribution rooms, substations, etc., which work together to complete exception handling.

[0123] Take an office building as an example. When the system monitors the power consumption curves of each room, it finds that the power consumption of Company A's office is much higher than that of other offices on the same floor, and higher than the daily level of Company A itself, and it is judged as an abnormality at the room level. The system retrieves the preset strategy library and matches the "Second Optimization Strategy for Room Level Abnormality", that is, power restriction is implemented for the room. Then, the system finds that the total power consumption of the 8th floor has a short-term spike at the building level, and it is judged as an abnormality at the building level. After matching the "Third Optimization Strategy" in the library, it is decided to implement time-sharing power supply to the public area of ​​the 8th floor to stagger the peak power consumption. Subsequently, the system found that the overall power load of the office building showed the characteristics of "unbalanced high and low floors, east and west wings", which is an abnormality at the building level. According to the "Fourth Optimization Strategy", the system generates a balanced scheduling instruction of "dispatching redundant power supply from 4-6 floors to 1-3 floors during peak hours, and redundant power supply from the east wing to the west wing". Finally, the system perceives that due to the peak power consumption in the nearby industrial park, the power grid load in the area where the office building is located has increased sharply, and the power supply is in urgent need of power supply. According to the "fifth optimization strategy" at the regional level, the system instructs the regional power distribution control center to dispatch a certain amount of power resources from the industrial park to the commercial area where the office buildings are located to alleviate the contradiction between supply and demand.

[0124] S208. Send the device optimization instruction to the control module of the abnormal device, so that the abnormal device executes the corresponding optimization instruction.

[0125] The control module refers to the embedded control unit inside the device, which is responsible for receiving external instructions and driving the actuator to perform corresponding actions. The optimization instruction refers to the control instruction sequence that adjusts the parameters or logic of the device to improve its operating status.

[0126] Specifically, the power consumption monitoring system first establishes a communication connection with the device through industrial interoperability protocols such as OPC UA (Open Production Control and Unified Architecture) and starts a command transmission session. Then, the system packages the optimized instruction set into lightweight data formats such as JSON, and clearly indicates the execution timing and execution conditions of each instruction. Next, the system transmits the packaged instructions to the control module on the device side in batches through the session channel. The control module monitors the channel in real time, performs syntax and semantic analysis on the received instructions, and converts them into register operations or IO port controls that can be recognized by the device hardware. Finally, the control module selectively triggers the corresponding execution actions according to the timing and conditions specified in the instructions to complete the optimization adjustment of the device status.

[0127] In some embodiments, the distribution and execution of optimization instructions to devices can be achieved in a variety of ways: Optionally, a publish-subscribe message transmission protocol such as MQTT is used to publish device instructions as JSON messages to the Topic (topic, used to define the channel and path for message transmission in the MQTT protocol) corresponding to the device; the control module obtains instructions by subscribing to the Topic, and uses state machine-based logic control to execute the action sequence. Optionally, a request-response interaction protocol such as HTTP is used to encapsulate the device instructions in the request parameters and send them to the Web API interface open to the control module; the control module obtains instructions by parsing the request parameters, and uses the interpreter mode to dynamically execute the instruction script. It is understandable that other industrial communication protocols and embedded control technologies can also be used to achieve closed-loop scheduling from upper-level optimization instructions to device-level execution actions, which is not limited here.

[0128] S209: Obtain a power supply line topology map of the abnormal floor, and identify other rooms located on the same power supply branch as the abnormal room based on the power supply line topology map.

[0129] The power supply line topology diagram refers to a connectivity diagram that reflects the internal power network structure of a floor. Nodes represent branch components such as switches and transformers, and edges represent wire connections. The power supply branch is used to represent a complete power supply path from the distribution box to the terminal load equipment.

[0130] Specifically, the power consumption monitoring system first obtains the CAD drawings of the floors from the building management system or the power supply and distribution GIS system, and extracts the layer information related to power supply. Then, the system uses the image recognition algorithm to locate and classify the icons of the power supply components and obtain their spatial coordinate information. Next, the system automatically generates a directed graph model that reflects the power supply topology of the floor based on the connection relationship of the wire graphics. On this basis, the system combines the location coordinates of the abnormal room, searches for the corresponding power supply branch in the topology map, and marks other power-consuming rooms that share the branch with it. Finally, the spatial distribution and topological path of these associated rooms are output.

[0131] In some embodiments, the identification of associated rooms on the power supply branch can be achieved in a variety of ways: optionally, a symbol recognition method based on template matching is used to extract the power supply components in the drawing, and then a vectorization algorithm is used to extract the wire connection topology; the shortest path search is used on the topology map to determine the power supply link of the abnormal room, and other room nodes on the same path are retrieved. Optionally, an object detection model based on machine learning is used to classify and locate drawing elements, and the room location attributes are annotated while constructing the power supply topology; the electrical similarity of the room nodes in the topology map is calculated using graph embedding representation learning technology, and the nodes with the highest similarity are clustered into rooms on the same branch. It is understandable that other graphics and image processing and topology data modeling and analysis technologies can also be used to achieve room-level decoupling and association impact analysis of complex floor power supply networks, which are not limited here.

[0132] S210. Calculate the proportion of the power load of the abnormal room to the total load of the abnormal floor.

[0133] The power load refers to the total power demand of all electrical equipment in the room per unit time, usually in kilowatts (kW). The power consumption ratio refers to the percentage weight of the load of a single room to the total load of the entire floor.

[0134] Specifically, the power consumption monitoring system first obtains the real-time power consumption monitoring data of the abnormal room and each device in it, and estimates its total load demand using data cleaning and feature extraction methods. At the same time, the system accumulates the power consumption components of the entire floor branch by branch in the power supply line topology diagram to calculate the total load benchmark. Then, the system divides the load of the abnormal room by the total load of the floor, and then multiplies it by 100% to obtain its power consumption ratio index. Next, the system determines whether the ratio exceeds the preset capacity threshold. If it exceeds, it means that the abnormal room has become a "large power consumer" on this floor, and its abnormal fluctuations will significantly affect the power supply quality of other rooms. Finally, the system outputs the ratio calculation result and triggers an alarm.

[0135] In some embodiments, the evaluation of the proportion of electricity consumption in a room can be achieved in a variety of ways: optionally, using the physical parameter estimation method, a load composition model is established based on factors such as the area of ​​the room, the rated power of the equipment, and the intensity of business operations; using the weighted average to aggregate the branch loads on each floor, the load proportion of the abnormal room in the spatial and electrical topology dimensions is calculated. Optionally, using the machine learning fitting method, the online estimation of load demand is achieved by training the nonlinear relationship between the room electricity consumption curve and the total electricity consumption of the floor; using the independent component analysis technique to decouple the floor electricity consumption curve and calculate the variance contribution of the abnormal room load signal. It is understandable that other load modeling, spatiotemporal data fusion, statistical signal analysis and other technologies can also be used to achieve room-level deconstruction of complex electricity consumption behaviors within the floor, which is not limited here.

[0136] S211. When the power consumption ratio exceeds a preset power consumption threshold, short-term power restriction is performed on other rooms to ensure power supply to the abnormal rooms.

[0137] The power consumption ratio refers to the proportion of the actual power consumption of the abnormal room to the total power consumption of the floor where it is located. The preset power consumption threshold represents the warning line of the power consumption ratio of the abnormal room. If the threshold is exceeded, the power consumption of the room is considered to be seriously abnormal and optimization measures need to be taken. Short-term power limitation refers to reducing the upper limit of the available power of other normal rooms within a certain period of time to free up power resources to ensure the power supply needs of the abnormal room.

[0138] This step is initiated when the power monitoring system detects an abnormal increase in power consumption in a room, causing its power consumption ratio to exceed the preset threshold. The sharp increase in power consumption in the abnormal room may squeeze out the power resources of other rooms on the same floor. In order to avoid large-scale undervoltage or power outages, emergency power limitation measures need to be taken for other rooms to replenish the excess power consumption of the abnormal room.

[0139] Specifically, the system first identifies the room with abnormal power consumption, calculates the proportion of its power consumption to the total power consumption on the same floor, and determines whether it exceeds the pre-set warning threshold. Once exceeded, the short-term power limitation process is immediately triggered. The system sends power limitation instructions to other rooms except the abnormal room, notifying them to control the power consumption within a relatively low upper limit in the short term. Non-essential electrical equipment in these rooms may be required to suspend operation for a period of time to reduce the total power consumption. At the same time, the saved power resources are preferentially allocated to rooms with abnormal power consumption to ensure the normal operation of their core equipment. When the power consumption ratio of the abnormal room falls below the threshold and is maintained for a certain period of time, the short-term power limitation measures are lifted and other rooms resume normal power supply mode.

[0140] In some embodiments, short-term power limitation can be achieved in a variety of ways: Optionally, the system can change the power supply configuration parameters of other rooms by controlling the smart meter, setting the maximum available power of these rooms to a lower value, so that they are forced to reduce power consumption. At the same time, the maximum available power parameter of the meter in the abnormal room is increased, allowing it to exceed the power consumption for a short period of time. After a period of time, the system restores the meter parameters to normal levels. Optionally, the system can achieve power limitation by directly controlling the electrical equipment in other rooms. The system issues a control instruction, requiring the non-core equipment such as air conditioners and lighting in these rooms to be temporarily shut down to reduce power consumption. At the same time, the electrical equipment in the abnormal room is instructed to keep running. After the abnormal situation is resolved, the instruction is issued again to allow the equipment in other rooms to resume operation.

[0141] It is understandable that in addition to adjusting meter parameters and directly controlling equipment, other power-limiting methods such as timed power outages and time-sharing power supply can also be used to achieve short-term power limitation, which are not listed in detail here. The system can flexibly select the best solution based on actual conditions, while ensuring power supply to abnormal rooms and minimizing the impact on other rooms.

[0142] In the embodiments of the present application, due to the use of a series of intelligent means such as full-dimensional topological modeling, precise scene perception, hierarchical abnormality diagnosis and collaborative optimization control, it is possible to comprehensively characterize the regional power consumption structure and behavior, timely discover local anomalies, and achieve global balance on the basis of giving priority to key needs. It effectively solves the problems of insufficient granularity of internal power consumption supervision and lack of effective emergency dispatch mechanism in key places such as hospitals, thereby realizing safe, reliable and efficient smart power supply and use.

[0143] The following describes the power consumption monitoring system in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , which is a schematic diagram of a physical device structure of the electricity consumption monitoring system in an embodiment of the present application.

[0144] It should be noted that Figure 3 The structure of the power consumption monitoring system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0145] like Figure 3 As shown, the power consumption monitoring system includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 to the random access memory (RAM) 303, such as executing the method described in the above embodiment. In RAM 303, various programs and data required for system operation are also stored. CPU 301, ROM 302 and RAM 303 are connected to each other through bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0146] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, a button switch, etc.; an output section 307 including a liquid crystal display (LCD) and an audio output device, an indicator light, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0147] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are performed.

[0148] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0149] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings.

[0150] Specifically, the electricity consumption monitoring system of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, the regional electricity consumption monitoring and control method provided in the above embodiment is implemented.

[0151] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the power consumption monitoring system described in the above embodiment; or may exist independently without being assembled into the power consumption monitoring system. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of the power consumption monitoring system, the power consumption monitoring system implements the monitoring and control method of regional power consumption provided in the above embodiment.

[0152] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0153] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0154] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. A method for monitoring and controlling regional electricity consumption, characterized in that: Applied to a power consumption monitoring system, the method comprises: Based on the regional power network structure data and user power settings, a multi-level power consumption model including device layer, room layer and floor layer is constructed; there are multiple floors in the floor layer, multiple rooms in the room layer, multiple devices in the device layer, one floor corresponds to multiple rooms, and one room corresponds to multiple devices; Acquire the associated feature data of the target room including the number of people in the room and the equipment list, and determine the real-time business scenario of the target room based on the associated feature data; the target room is a room in the room layer; Based on the historical power consumption data of multiple similar rooms in the room layer on the multi-level power consumption model, determine the power consumption data boundary of the target room in the real-time business scenario; the equipment corresponding to the similar rooms is the same as that of the target room; When it is detected that the real-time power consumption data corresponding to the target room exceeds the power consumption data boundary, the target room is regarded as an abnormal room, and the abnormal equipment and abnormal floor corresponding to the abnormal room are determined based on the multi-level power consumption model; After determining the abnormal characteristics of the abnormal device according to the real-time business scenario and the real-time power consumption data, generating a device optimization instruction based on the abnormal characteristics and sending the device optimization instruction to the abnormal device; the device optimization instruction includes adjusting operating parameters or temporarily cutting off power supply; Based on the abnormal feature and the proportion of electricity consumption of the abnormal room on the abnormal floor, power supply scheduling of multiple room layers on the abnormal floor is controlled.

2. The method according to claim 1, characterized in that The step of obtaining the associated characteristic data of the target room including the number of people in the room and the equipment list, and determining the real-time business scenario of the target room based on the associated characteristic data, specifically includes: Acquire a real-time image of the target room, and determine the number of people in the room based on the real-time image; Determine the equipment list of the target room in combination with the real-time image and user settings; the equipment list includes equipment model information, power parameters and quantity information of various electrical equipment configured in the target room; Taking the number of people in the room and the equipment list as associated characteristic data, determining a service adaptation set according to the target room and the associated characteristic data; Inputting the associated feature data into a business scenario recognition model corresponding to the business adaptation set to obtain a real-time business scenario; When the number of people in the room or the equipment list changes, the changed associated feature data is re-acquired, and the real-time business scenario of the target room is dynamically updated.

3. The method according to claim 2, characterized in that The method of using the number of people in the room and the equipment list as associated characteristic data and determining a service adaptation set according to the target room and the associated characteristic data specifically includes: Presetting a plurality of corresponding adaptation service scenarios according to the room type and room purpose of the target room; In the adaptation service scenario, a first service subset having a number matching degree higher than a first threshold is screened out according to the number of people in the room; Obtain a historical equipment list of the room, compare the historical equipment list with the equipment list, and obtain a newly added equipment list and a reduced equipment list; If both the newly added equipment list and the reduced equipment list are empty, obtaining the historical business scene of the target room at the previous moment, and filtering out a second business subset from the first business subset according to the historical business scene, the second business subset having a deviation from the historical business scene less than a second threshold; If the newly added device list is not empty, calculating the matching degree between the devices in the newly added device list and the standard devices corresponding to each business scenario in the first business subset, and filtering out a second business subset from the first business subset with a matching degree greater than a third threshold according to the matching degree; If the reduced equipment list is not empty, removing the business scenarios including the devices in the reduced equipment list from the standard devices corresponding to each business scenario in the first business subset, to obtain a second business subset; The second service subset is determined as the service adaptation set of the target room.

4. The method according to claim 1, characterized in that: After determining the abnormal characteristics of the abnormal device according to the real-time business scenario and the real-time power consumption data, generating a device optimization instruction based on the abnormal characteristics and sending the device optimization instruction to the abnormal device specifically includes: Determine various power usage parameters of the abnormal device in multiple power usage time periods according to the real-time business scenario and the real-time power usage data; the power usage parameters include voltage, current, power factor and harmonics; Determine, based on the various power usage parameters, abnormal characteristics of the abnormal equipment, including equipment performance degradation, power usage risk level, and energy efficiency level; According to the abnormal characteristics, a device optimization strategy is matched in a preset strategy library, and a corresponding device optimization instruction is generated based on the device optimization strategy; the device optimization instruction includes one or more of the following: adjusting the operating parameters of the device, updating the control logic of the device, or cutting off the power supply of the device; The device optimization instruction is sent to the control module of the abnormal device, so that the abnormal device executes the corresponding optimization instruction.

5. The method according to claim 4, characterized in that The multi-level electricity consumption model also includes a building layer and a regional layer. According to the abnormal characteristics, the device optimization strategy is matched in the preset strategy library, and the corresponding device optimization instruction is generated based on the device optimization strategy, which specifically includes: In the preset strategy library, a first optimization strategy for device layer abnormalities, a second optimization strategy for room layer abnormalities, a third optimization strategy for floor layer abnormalities, a fourth optimization strategy for building layer abnormalities, and a fifth optimization strategy for area layer abnormalities are obtained according to the abnormal feature matching; the first optimization strategy is used to indicate adjustment of operating parameters of abnormal equipment or temporary cutting off of its power supply; the second optimization strategy is used to indicate power limitation or timed power off of abnormal equipment in the room layer; the third optimization strategy is used to indicate total control or time-sharing power supply for abnormal rooms in the floor layer; the fourth optimization strategy is used to indicate power load balancing between floors in the building layer; the fifth optimization strategy is used to indicate power dispatching for each building in the area layer; Determining a target optimization strategy among the first optimization strategy, the second optimization strategy, the third optimization strategy, the fourth optimization strategy, and the fifth optimization strategy according to the abnormal level corresponding to the abnormal feature; According to the target optimization strategy, a device optimization instruction of a corresponding level is generated, and the device optimization instruction is sent to the node device of the corresponding level.

6. The method according to claim 1, characterized in that The controlling the power supply scheduling of multiple room layers on the abnormal floor based on the abnormal feature and the proportion of power consumption of the abnormal room on the abnormal floor specifically includes: Acquire a power supply line topology map of the abnormal floor, and identify other rooms located on the same power supply branch as the abnormal room based on the power supply line topology map; Calculate the proportion of the power load of the abnormal room to the total load of the abnormal floor; When the power consumption ratio exceeds a preset power consumption threshold, short-term power restriction is performed on the other rooms to ensure power supply to the abnormal room.

7. The method according to claim 1, characterized in that The multi-level power consumption model including the equipment layer, room layer and floor layer is constructed based on the regional power network structure data and the user's power consumption settings, specifically including: Receive and parse regional power network structure data, and determine the topological connection relationship between device nodes, room nodes, floor nodes and nodes in the regional power network structure data; Accessing the electricity consumption information database to extract attribute information of each node; the attribute information includes rated parameters of the equipment, area of ​​the room, functional attributes and vertical position information of the floor; Generate a multi-level power consumption topology network according to the topological connection relationship and the attribute information; the multi-level power consumption topology network is a hierarchical structure including a device layer, a room layer and a floor layer; Receiving power usage settings input by a user in a user terminal, and mapping and marking the power usage settings as constraints in the multi-level power usage topology network; the power usage settings include power usage time periods of each room, rated power of each device, and power usage priority parameters; Based on the annotated multi-level electricity consumption topology network, a multi-level electricity consumption model is constructed.

8. A power consumption monitoring system, characterized in that: The power consumption monitoring system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the power consumption monitoring system to execute the method described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instruction is executed on the power consumption monitoring system, the power consumption monitoring system is caused to execute the method as claimed in any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product runs on a power consumption monitoring system, the power consumption monitoring system is enabled to execute the method according to any one of claims 1 to 7.

Images