Power supply abnormity identification and self-adaptive scheduling method and system

Abstract

The invention relates to a power supply abnormity identification and self-adaptive scheduling method and system, and the method comprises the steps: obtaining a CAD power design drawing, constructing a power system topology model based on a graph database, constructing the power system topology model based on the graph database, and carrying out the power supply abnormity identification and self-adaptive scheduling. According to the method, multi-source sensor data, emergency lighting and image recognition signals are fused to carry out abnormal cross validation and accurate positioning, abnormal retrieval area determination and power supply abnormal recognition are realized, and on the basis of equipment real-time operation margin evaluation, cable current-carrying capacity, switch margin, electrical accessibility, space accessibility and space fire partition are taken as constraints, and power supply abnormal recognition is realized. And a self-adaptive scheduling strategy is generated through a multi-objective optimization function, so that rapid fault isolation, load loss minimization and power supply reliability maximization are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system automation and smart grid, and particularly relates to a power supply abnormality identification and adaptive scheduling method and system. BACKGROUND

[0002] With the increasing demand for power supply reliability in modern industry, commercial buildings and data centers, the safe and stable operation of power supply and distribution systems is crucial. The traditional power supply abnormality processing method has many drawbacks.

[0003] In the prior art, an invention patent with publication number CN121076762A discloses a power emergency decision method, which determines the abnormal point of failure according to the real-time data of each device node; determines the search area centered on the abnormal point in the electrical interval unit where the abnormal point is located; determines the available device nodes in the search area through the electrical reachable node set and the spatial reachable node to realize the synchronous search of power failure point spatial positioning and electrical connectivity, solves the problem of spatial and electrical connectivity data fragmentation, and improves the search efficiency of device nodes, but this method mainly focuses on searching for standby switching power supply nodes, lacks root cause analysis of faults and safety constraints related to fire compartment in the power supply recovery process.

[0004] An invention patent with publication number CN119538153A discloses a power supply scheduling abnormality detection system and method, a data processing and analysis module configured to process and analyze the parameter data using a deep learning algorithm to identify possible abnormal conditions in the power system; an abnormality detection and identification module configured to detect and identify the abnormal conditions using data mining techniques to determine the type and severity of the abnormal conditions, but this method has a large amount of data collection and analysis, lacks fault confirmation verification, and does not provide a power supply recovery scheduling strategy under abnormal conditions.

[0005] It can be seen that the traditional power supply abnormality identification and scheduling still has the following problems:

[0006] Serious information island phenomenon: power monitoring systems (SCADA), fire protection systems (such as emergency lighting start), security systems (video monitoring), etc. are usually operated independently. The data of each system cannot be effectively integrated, resulting in that the abnormal signal of a single system may be misjudged or ignored, lacking multi-dimensional cross verification, and the identification accuracy is not high.

[0007] The emergency treatment strategy is rigid: the recovery power supply strategy after failure is a pre-set fixed scheme (FA-feeder automation), which cannot be dynamically adjusted according to real-time equipment working conditions (such as cable temperature, switch residual life), load importance and space environment (such as fire compartment). This may lead to a non-optimal recovery path, or even cause secondary failure, for example, transferring load to a cable that has reached the load limit.

[0008] Lack of predictability and adaptability: existing systems are mostly passive responses, that is, measures are taken only after the failure occurs. There is a lack of real-time assessment of real-time operating margin and design margin of equipment operation, which cannot provide early warning before failure occurs, nor can it adaptively schedule in a globally optimal and minimum equipment loss manner after failure.

[0009] Therefore, there is an urgent need for a new method that can break down information barriers, achieve accurate and rapid identification, and intelligently and adaptively schedule based on real-time system state, to comprehensively improve the intelligent level and power supply reliability of the power supply and distribution system. SUMMARY

[0010] The purpose of the present application is to overcome the shortcomings of the prior art and provide a power supply anomaly identification and adaptive scheduling method and system. The method constructs a digital twin topology model synchronized with the physical system in real time, fuses multi-source heterogeneous data for accurate identification and positioning of anomalies, and on this basis, performs adaptive scheduling based on real-time state of equipment and multi-dimensional constraints, achieving rapid fault isolation, minimum load loss and maximization of power supply reliability.

[0011] To solve the above technical problems, the first aspect of the present application provides a power supply anomaly identification and adaptive scheduling method, characterized in that it comprises the following steps:

[0012] S1: Obtain a CAD power design drawing, and construct an electric power system topology model based on a graph database, wherein the electric power system topology model comprises a power supply path, a power switch specification, a cable diameter specification, and a sensor arrangement position;

[0013] S2: Correct the electric power system topology model based on the sensor collected parameters;

[0014] S3: Associate an emergency lighting module and / or an image recognition module in a monitoring system, determine an abnormal search area based on abnormal signals of data collected by the sensor, the emergency lighting module and the image recognition module, perform power supply anomaly identification on the abnormal search area, and the identification result comprises power supply short circuit, overload and step-out tripping;

[0015] S4: Based on the power supply abnormality identification result, the real-time operation margin and the design margin of the power switch device are evaluated, a multi-objective optimization function is constructed with cable real-time load flow and switch margin, electrical accessibility, spatial accessibility and space fire compartment as constraints, and a power supply adaptive scheduling strategy is output.

[0016] Optionally, the step S1 specifically comprises:

[0017] S1.1: Analyzing the CAD power design drawing, identifying electrical element symbols and connection lines, and extracting element specification parameters and spatial position information;

[0018] S1.2: In the graph database, electrical elements are created as nodes, power supply paths are created as edges, and the specification parameters and spatial position information are taken as node and edge attributes;

[0019] S1.3: The sensors include residual current monitoring sensors, voltage / current sensors, and / or temperature sensors, and the monitoring parameter types and arrangement position information of the sensors are associated with the corresponding node or edge attributes.

[0020] Optionally, the step S2 specifically comprises:

[0021] Real-time collection of monitoring data of each sensor, association of the monitoring data with the corresponding nodes or edges in the graph database, dynamic updating of the real-time state attributes of the nodes or edges, and construction of a dynamic digital twin model synchronized with the physical power grid.

[0022] Optionally, the step S3 specifically comprises:

[0023] S3.1: When receiving any one of a sensor data abnormality, an emergency lighting module start or an image recognition module identifying an abnormal visual feature, a preliminary abnormality search area is circled in the power system topology model;

[0024] S3.2: In the preliminary abnormality search area, cross-verification is performed by querying abnormal signals of at least another data source, if the verification is passed, the preliminary abnormality search area is confirmed as an abnormality search area, and the fault point is accurately located using multi-source information.

[0025] Optionally, the multi-objective optimization function in the step S4 is:

[0026] ,

[0027] Wherein: , , is a weight coefficient, is a load loss cost, is a switch operation cost, Benefits for power supply reliability;

[0028]

[0029]

[0030]

[0031] in, It is a load node in the system. For the set of load nodes, For load Importance weight, For load Power supply state variables, Power supply status. The system is in a power-off state. It's a switch In the initial state before the failure occurs It is a load Reliability score.

[0032] Optionally, the constraints in step S4 include:

[0033]

[0034]

[0035]

[0036]

[0037]

[0038]

[0039]

[0040] in, Power to load The set of switches on possible paths, For power supply collection, Indicates the number of power nodes. For flow through the switch Real-time current, For switch Rated current, For the first Real-time current of each device For the first The rated current carrying capacity of the power supply cable for each device. This is a temperature correction factor. electrical operability of the switch physical operability of the switch whether the load is powered by the power supply power supply accessibility for a fire compartment.

[0041] In a second aspect, the present application also provides a power supply anomaly identification and adaptive scheduling system, comprising:

[0042] a model construction module, configured to acquire a CAD power design drawing and construct a power system topology model based on a graph database;

[0043] a model correction module, configured to correct the power system topology model in real time based on sensor-acquired parameters to form a dynamic digital twin model;

[0044] an anomaly identification module, configured to associate an emergency lighting module and / or an image recognition module in a monitoring system, perform anomaly searching and fault type identification through multi-source data fusion;

[0045] an adaptive scheduling module, configured to evaluate equipment real-time operation margin based on a fault identification result and the dynamic digital twin model, and output a power supply adaptive scheduling strategy in a multi-objective optimization function.

[0046] In a third aspect, the present application also provides an electronic device, comprising:

[0047] one or more processors;

[0048] a memory; and one or more programs stored in the memory, which include instructions for performing the power supply anomaly identification and adaptive scheduling method described above.

[0049] In a fourth aspect, the present application also provides a computer-readable storage medium, comprising one or more programs for execution by one or more processors of an electronic device, which include instructions for performing the power supply anomaly identification and adaptive scheduling method described above.

[0050] Compared with the prior art, the present application has the following remarkable beneficial effects:

[0051] 1. High identification accuracy and fast speed: cross verification is performed through fusion of electrical, visual, and state multi-source information, which greatly improves the accuracy of anomaly identification, effectively avoids the problem of false positives or false negatives of a single sensor, and can quickly and accurately locate the fault point.

[0052] ​​​2. Intelligent and adaptive scheduling: Abandoning rigid pre-planned recovery, a multi-objective optimization model is adopted, considering load loss, operation cost, power supply reliability and equipment health status, which can generate the optimal scheduling strategy adapting to the current working condition, realizing the transformation from "passive response" to "active intelligent decision".

[0053] 3. Improving system safety and economy: By considering the real-time operation margin and design allowance of equipment, secondary damage to fragile equipment during power restoration is avoided, prolonging the service life of equipment. By minimizing the number of switch operations and load loss, the operation and maintenance cost and power loss are reduced.

[0054] 4. Global perspective and multi-dimensional constraints: The multi-dimensional constraints such as electrical topology, equipment state, physical space, fire safety, etc. are included in the decision-making process, so that the scheduling scheme is not only technically feasible, but also compliant and reliable in physical space and safety standards. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 A power supply abnormality identification and adaptive scheduling method flow chart is provided for the embodiments of the present application;

[0056] Figure 2 A power system topology model schematic diagram is provided for the embodiments of the present application;

[0057] Figure 3 An adaptive scheduling strategy flow chart is provided for the embodiments of the present application. DETAILED DESCRIPTION

[0058] Obviously, many modifications and variations of the present application can be made by those skilled in the art based on the purpose of the present application, which belong to the protection scope of the present application.

[0059] Those skilled in the art can understand that, unless specifically stated, the singular forms "a", "an" and "the" used herein also include the plural forms. It should be further understood that the use of the phrase "comprising" in the specification of the present application means that the features, integers, steps, operations, elements and / or components exist, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their combinations. It should be understood that when an element, component is said to be "connected" to another element or component, it can be directly connected to the other element or component, or there can be intermediate elements or components. The phrase "and / or" used herein includes any one of the associated listed items and all combinations thereof.

[0060] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0061] The technical solutions of the present application will be described in more detail below with reference to the drawings and specific embodiments.

[0062] With reference to Figure 1 The power supply abnormality identification and adaptive scheduling method provided in the embodiment includes the following steps.

[0063] S1: Obtain a CAD power design drawing, and construct a power system topology model based on a graph database, wherein the power system topology model includes a power supply path, power switch specifications, cable diameter specifications, and sensor arrangement positions.

[0064] S2: Correct the power system topology model based on sensor collected parameters.

[0065] S3: Associate an emergency lighting module and / or an image recognition module in a monitoring system, determine an abnormal search area based on abnormal signals of data collected by the sensors, the emergency lighting module and the image recognition module, identify power supply abnormalities in the abnormal search area, and the identification result includes power supply short circuit, overload and step tripping.

[0066] S4: Based on the power supply abnormality identification result, evaluate the real-time operation margin and design margin of the power switch device, construct a multi-objective optimization function with cable real-time current-carrying capacity and switch margin, electrical accessibility, spatial accessibility and space fire compartment as constraints, and output a power supply adaptive scheduling strategy.

[0067] In an optional embodiment, step S1: constructing a power system topology model based on a graph database specifically includes:

[0068] Obtain a CAD power design drawing, and construct a power system topology model based on a graph database by analyzing element information, connection relationships and attribute parameters in the drawing. The power system topology model takes power devices as nodes and cables as edges, and the attributes include:

[0069] Power supply path: complete electrical path from power supply point to each load node.

[0070] Power switch specifications: including model, rated current, rated voltage, trip curve, etc. of circuit breakers and disconnectors.

[0071] Cable size: including the model of the cable, cross-sectional area, rated current carrying capacity, length, impedance, etc.

[0072] Sensor arrangement position: the sensors include residual current monitoring sensors, voltage / current sensors, and / or temperature sensors. The specific node positions of the residual current monitoring sensors, voltage / current sensors, and temperature sensors in the topology model are associated.

[0073] In an optional embodiment, taking a high-rise commercial complex as an example, its 10kV / 0.4kV power distribution system contains two transformers (T1, T2), multi-face low-voltage distribution cabinets (such as AA1, AA2), and hundreds of feeder circuits.

[0074] First, obtain its full set of CAD design drawings (in DWG format), use the electrical drawing parsing engine, and perform the following operations:

[0075] Layer and block identification: identify and separate the “circuit breaker layer”, “cable layer”, “load layer”, and “annotation layer”. Parse the dynamic blocks in the drawing, such as the circuit breaker block, and extract its attribute information including the circuit breaker specification model, rated current In, and trip curve.

[0076] Topology relationship extraction: by analyzing the start and end points of the cable connection line, the connection relationship between electrical elements is automatically established.

[0077] Model construction: in a graph database (such as Neo4j), each electrical device (transformer, circuit breaker, load) is created as a node, and each cable is created as an edge.

[0078] Node attribute example: {id: “QF2”, name: “AA1 main incoming line”, type: “ACB”, rated_current: 2500A, position: (x1, y1, z1), is_remote_controllable: true}

[0079] Edge attribute example: {id: “Cable-E3”, type: “YJV”, cores: “4x240+1x120”, rated_amps: 380A, length: 55m, derating_factor: 0.8}

[0080] Sensor association: the residual current sensor (RCD) S_RCD_QF7 is associated as an attribute to the circuit breaker node QF7 it monitors; the cable temperature sensor S_Temp_E5 is associated to the cable edge E5.

[0081] In an optional embodiment, refer toFigure 2 In the constructed topology model, node N1 represents a 10kV transformer, N2, N3, N4 represent the main circuit breakers in the low-voltage distribution cabinet, N5-N10 represent the molded case circuit breakers of each feeder circuit, and L1-L8 represent the electrical loads. Edges E1-E8 represent the cables connecting these nodes. Each node and edge stores the specification parameters extracted from the CAD drawings, such as the rated current of N2 being 2500A, the cable model of E3 being YJV-4x240+1x120, etc. At the same time, information such as the RCD sensors installed in N6, N7 circuits (S_RCD6, S_RCD7) and the sensors monitoring the temperatures of cables E4, E5 (S_Temp4, S_Temp5) are associated with the corresponding nodes and edges.

[0082] In an optional embodiment, step S2: correcting the power system topology model based on sensor data:

[0083] The system accesses the data stream of the Internet of Things platform in real time through protocols such as MQTT. For example, the data {sensor_id: "S_RCD7", value: 305A, timestamp:...} is received. The system finds the associated node N7 in the graph database according to the ID "S_RCD7" and updates its attribute "real-time current" to 305A. Similarly, when {sensor_id: "S_Temp4", value: 65°C,...} is received, the "real-time temperature" of edge E4 is updated to 65°C. Through this continuous data synchronization, the graph database model becomes a dynamic digital twin that reflects the real power grid operation.

[0084] In an optional embodiment, step S3: anomaly retrieval and fault identification of multi-source data fusion specifically includes:

[0085] Preliminary anomaly retrieval: when an anomaly signal from any source of sensors, emergency lighting modules, or image recognition modules is received, preliminary anomaly retrieval is triggered, for example:

[0086] Sensor signal: residual current sensor alarm, sudden increase / decrease in current, voltage drop, temperature overrun;

[0087] Emergency lighting module signal: concentrated lighting of emergency lighting in abnormal state;

[0088] Image recognition module signal: identification of abnormal visual features such as arc, spark, smoke, etc.

[0089] According to the anomaly signal, one or more possible anomaly retrieval areas are preliminarily circled in the power system topology model.

[0090] Anomaly confirmation and identification: In the preliminary search area, further cross-verification is performed by fusing data from at least another source. If the verification is passed, the area is determined as the final anomaly search area. By analyzing the time series and logical relationship of various data in this area, the specific fault type is identified, including:

[0091] Power supply short circuit: characterized by instantaneous maximum current, voltage drop, and possible arc flash.

[0092] Power supply overload: characterized by continuous current higher than rated value, slow temperature rise, and possible delayed tripping of switch.

[0093] Power supply step-out trip: characterized by the upper switch tripping, but its direct lower switch does not act according to the protection logic. It is determined by analyzing the hierarchical relationship of switch action in the topology model.

[0094] In an optional embodiment, it is assumed that at 14:00, the system simultaneously receives multiple anomaly signals.

[0095] Electrical signal: The current transformer CT_QF7 detects that the current jumps from 155A to 3200A in an instant, and the voltage monitoring point V_L5 voltage drops to 50V.

[0096] Fire signal: The fire system API transmits the event: {area: "5F-East Wing", event: "Emergency Lighting Activated"}.

[0097] Security signal: The AI video analysis module of the camera Cam_5F_E01 located in the east corridor of 5F generates an alarm: {camera_id: "Cam_5F_E01", event: "Arc Flash Detected", confidence: 98%}.

[0098] The system performs the following identification process:

[0099] S3.1: Preliminary area determination: The source of the current mutation is the circuit breaker QF7. The system immediately locates QF7 in the topology model and encircles all devices downstream of QF7 (including cables E8, E9 and loads L5, L6) as the "preliminary anomaly search area".

[0100] S3.2: Area confirmation and precise positioning: The system queries other data sources in the preliminary area. It finds that:

[0101] The emergency lighting of the fire zone "5F-East Wing" corresponding to this area has been started.

[0102] The camera Cam_5F_E01 in this area detects arc light.

[0103] With the strong correlation signal confirmation of the two independent systems of fire protection and security, the system upgrades the "preliminary abnormal search area" to the "abnormal search area". By analyzing the pixel coordinates of the arc light in the image, combined with the known position and attitude of the camera, the system accurately locates the fault point to the terminal box near the load L5.

[0104] S3.3: Fault type identification:

[0105] Scenario one: short circuit fault identification

[0106] At T=0, the current of the incoming switch QF7 of PDU-A01 in column A cabinet jumps from 155A to 3200A in an instant, and the voltage monitoring point V_L5 voltage drops to 50V, preliminarily determining the fault area as PDU-A01 and its downstream. The system simultaneously queries the image recognition result of the monitoring camera Cam_5F_E01 covering the PDU-A01 area. The image recognition module reports the detection of strong arc flash at T=0.1 seconds. At the same time, the emergency lighting system reports that the emergency lights in this area are on due to the loss of mains power.

[0107] The system queries the event log including current mutation, voltage drop, arc flash, and emergency lighting start, and the four evidence chains completely coincide, and the system finally confirms that a short circuit fault occurs in the PDU-A01 area. Short circuit judgment: the current instantaneously increases to far exceed the rated value (3200A >> 155A), accompanied by voltage drop, which meets the characteristics of metallic short circuit.

[0108] Scenario two: overstep tripping fault identification

[0109] The system queries the event log and finds that the protection device of QF7 (molded case circuit breaker, rated 160A) does not send a trip signal, but its upper switch QF3 (frame circuit breaker, rated 630A) trips about 150ms after the short circuit occurs. Combined with the relationship that QF3 is the upper switch of QF7 in the topology model, the system determines that "a short circuit occurs at load L5, causing the upper switch QF3 to trip overstep".

[0110] In an optional embodiment, referring to Figure 3 , in step S4, a multi-objective optimization function is constructed with real-time cable current-carrying capacity and switch margin, electrical accessibility, spatial accessibility, and space fire compartment as constraints. The multi-objective optimization function is a weighted function with the objectives of minimizing load loss, minimizing switch operation times, and maximizing power supply reliability. The output is a power supply adaptive scheduling strategy, which specifically includes:

[0111] S4.1: Constructing a multi-objective optimization function:

[0112] The objective of this optimization model is to generate a power adaptive scheduling strategy to quickly restore power supply after power supply anomalies occur, while considering the operation cost and improving power supply reliability. This objective is achieved through a multi-objective weighted function, which includes the following three sub-objectives:

[0113] Minimize load loss: restore power supply to important loads as much as possible;

[0114] Minimize switch operation times: reduce equipment wear and tear and operation time;

[0115] Maximize power supply reliability: ensure that the restored power supply path has a high degree of health.

[0116] Let the set of switches in the power system be Define binary decision variables:

[0117] ,

[0118] where is the number of switches in the system.

[0119] The objective function F is a minimization problem, which is in the form:

[0120] ,

[0121] where: , , are the weight coefficients of the three sub-objectives, which are set by the operation and maintenance personnel according to the current operation and maintenance strategy (for example, in extreme weather, the weights of and can be increased to prioritize power supply), and the weight coefficients satisfy: , .

[0122] Load loss cost represents the total amount of power supply that has not been restored, which is a key indicator of the scope of power outage:

[0123]

[0124] where: is the load node in the system, is the set of load nodes, is the importance weight of the load (for example, a critical server =1, and ordinary lighting =0.2), is the power supply state variable of the load , is the power supply state, Loss of power state.

[0125] Switching operation cost Total switching operations required to execute the restoration strategy to reduce equipment wear and operation time.

[0126]

[0127] wherein, is the switching Initial state (0 or 1) before the fault occurs.

[0128] Power supply reliability gain The reliability level of the system after power restoration, which takes into account the number of critical loads and the health of the path to restore power to calculate:

[0129]

[0130] wherein, is the load Reliability score.

[0131] The importance of its importance is taken into account when calculating and the health of the path that supplies power to it, which can be determined by the real-time current carrying capacity margin of all cables on the path and the real-time margin of the switch:

[0132] ,

[0133] wherein, is the load Real-time margin ratio of the device (cable or switch) on the power supply path;

[0134] Real-time margin ratio of the cable: ,

[0135] Real-time margin ratio of the switch: ,

[0136] wherein, is the real-time current of the device, is the rated current carrying capacity of the power supply cable of the device, is the rated current of the device, is the temperature correction coefficient (can be referred to for value according to IEC 60364 standard).

[0137] S4.2: Set multiple constraints, and the optimization process must meet the corresponding physical and operational constraints:

[0138] a) Topology and radial operation constraints

[0139] To ensure the secure and stable operation of the power grid, the restored network must be radial (no loop) and each load can be supplied by at most one source:

[0140]

[0141] where, is the set of switches on the possible path from the source to the load , this constraint states that for each load , the sum of the states of all switches on the path connecting it to the source must be equal to its supply state ;

[0142] Loop-free constraint:

[0143]

[0144] where, is the set of sources, denotes the number of source nodes.

[0145] b) Electrical safety constraints

[0146] Real-time current carrying capacity of cables constraint: the real-time current flowing through any cable cannot exceed its safe current carrying capacity, which can be modified according to the real-time temperature:

[0147]

[0148] where, is the set of cables.

[0149] Switching device margin constraint: the real-time current flowing through any switch cannot exceed its rated current:

[0150] ,

[0151] where, is the real-time current flowing through switch , is the rated current of switch .

[0152] c) Spatial and environmental constraints

[0153] Electrical reachability constraint: when a fault occurs in a certain area, such as a short circuit fault, a fire fault, etc. fault scenario, it must be ensured that the fault point is effectively electrically isolated to prevent the fault from expanding. Define a binary parameter to represent the electrical operability of switch :

[0154] This parameter is directly determined by the power supply anomaly identification result in step S3, and the constraint condition is:

[0155] ,

[0156] This constraint ensures that switches used to isolate faults must remain open, which is a fundamental requirement for safe electrical operation and, together with topology constraints, determines the network's connectivity.

[0157] Spatial accessibility constraints: Define binary parameters based on real-time environmental data provided by fire alarm systems, security systems, or inspection robots. Indicates switch Physical operability:

[0158] ,

[0159] ;

[0160] Fire compartment constraints: Based on building information model or fire protection design drawings, buildings are divided into different fire compartments. To prevent the spread of fire, electrical connections between different fire compartments are usually restricted. Constraints include that general loads are not allowed to be powered across fire compartments, and fire protection loads and particularly important loads among the primary loads are not allowed to be powered across fire compartments.

[0161] Let the set of all fire compartments within the building be . For each load node and each power node Define its fire compartment:

[0162] :load The fire compartment it is located in;

[0163] :power supply The fire compartment it is located in;

[0164] Define the power supply accessibility of fire compartments:

[0165] ,

[0166] Fire compartment constraints for space:

[0167] ,

[0168] in, Indicates load Is it powered by a power source? Power supply, only when the power source Zone where the load is located When the power supply permission of the zone where the load is located is "allowed", the variable is 1.

[0169] S4.2: Solve the multi-objective optimization function, and output the adaptive scheduling strategy

[0170] The above model constitutes a mixed integer linear programming (MILP) problem. It can be solved by using mature optimization solvers such as Gurobi, CPLEX, or open source SCIP, and the present application does not make specific limitations. The solver inputs the objective function, all constraints and real-time parameters, and outputs a set of optimal switch state combinations. *}.

[0171] The *} output by the solver is the optimal adaptive scheduling strategy, which is converted by the system into a specific sequence of operation instructions (such as "open switch A, close switch B"), and is issued to the field power monitoring system (SCADA) or intelligent switch actuator to automatically complete power restoration.

[0172] In an optional embodiment, the fault event is a short circuit at L5, which causes QF3 to trip, and the lost loads are: L5 (primary load), L6 (secondary load), L7 (secondary load), and L8 (tertiary load).

[0173] The system presets the optimization target, and dynamically adjusts the weight according to the current priority of "important load preservation". Based on the priority of load loss, the operation cost is second, and the reliability is supplemented, and the reconfiguration is performed:

[0174] = 0.7, = 0.2, = 0.1.

[0175] Load loss cost : Weighted load loss is adopted, and the load importance weights of loads L5, L6, L7 and L8 are:

[0176] w_{imp,L5} = 1 (primary load, key server),

[0177] w_{imp,L6} = 0.5 (secondary load, important equipment),

[0178] w_{imp,L7} = 0.5 (secondary load, important equipment),

[0179] w_{imp,L8} = 0.1 (tertiary load, ordinary lighting).

[0180] Load information is as follows:

[0181]

[0182] Based on real-time data, calculate switch / cable equipment margin ratio:

[0183] Cable E2: I_rated = 2000A, I_real = 1750A, k_temp = 1 (temperature normal), δ_E2 = 1 - 1750 / 2000 = 0.125,

[0184] Switch QF2: I_rated = 2500A, I_real = 1800A, δ_QF2 = 1 - 1800 / 2500 =0.28;

[0185] Switch QF9 (T2 system): I_rated = 200A, I_real = 30A, δ_QF9 = 1 - 30 / 200 =0.85.

[0186] Electrical constraints: After any dispatch scheme is executed, real-time current carrying capacity of all lines ≤ rated current carrying capacity x derating factor.

[0187] Equipment constraints: Real-time operating margin of all switches ≤ 90%, leaving 10% emergency margin;

[0188] Topology constraints: Switches operated must be remotely controllable;

[0189] Space constraints: Assume QF3 is located in a power distribution room temporarily locked due to smoke alarm, personnel cannot access, but remote operation is not affected;

[0190] Safety constraints: 5F east wing and west wing belong to different fire zones. Regulations require that normal power supply should not cross fire zones unless fireproof cables are used.

[0191] Model solution and output adaptive dispatch strategy: The system searches all possible recovery schemes based on the topology model and evaluates them.

[0192] Scheme A: Restore L6, L7, L8 through tie switch QF_Tie

[0193] Operation: Keep QF3 open, close QF_Tie (assume initial state y_QF_Tie^0 = 0, y_QF3^0 =1).

[0194] Calculation: = |y_QF_Tie - y_QF_Tie^0| + |y_QF3 - y_QF3^0| = |1 - 0|+ |0 - 1| = 2.

[0195] Calculation: L5 only de-energized (x_L5=0, x_L6=1, x_L7=1, x_L8=1):

[0196] = w_{imp,L5}·(1 - x_L5) + w_{imp,L6}·(1 - x_L6) + w_{imp,L7}·(1 - x_L7) + w_{imp,L8}·(1 - x_L8) = 1·1 + 0.5·0 + 0.5·0 + 0.1·0 = 1.

[0197] Calculation: L6, L7, L8 restored, path health based on margin ratio:

[0198] L6 path: QF2 → QF_Tie, δ_min = min(δ_QF2, δ_QF_Tie) = min(0.28, 0.9) =0.28.

[0199] r_L6 = w_{imp,L6} · δ_min =0. 5 · 0.28 = 0.14.

[0200] L7 path: same as L6, r_L7 = 0.5 · 0.28 = 0.14.

[0201] L8 path: same as L6, r_L8 =0.1 · 0.28 = 0.028.

[0202] = r_L6·x_L6 + r_L7·x_L7 + r_L8·x_L8 = 0.14·1 + 0.14·1 +0.028·1 = 0.308.

[0203] F calculation: F = 0.7·1 + 0.2·2 - 0.1·0.308 = 7 + 0.4 - 0.308 = 1.0692.

[0204] Constraint check: new load current 380A, cable E2 current increased to 2130A > 2000A (violation of electrical constraint), scheme not feasible.

[0205] Scheme B: partial restoration, L8 disconnected

[0206] Operation: Close QF_Tie, open QF8 (isolate L8).

[0207] Calculation: = |y_QF_Tie - y_QF_Tie^0| + |y_QF8 - y_QF8^0| = |1 - 0| + |0 - 1| = 2.

[0208] Calculation: L5 and L8 de-energized (x_L5=0, x_L6=1, x_L7=1, x_L8=0):

[0209] = 1 · 1 + 0.5 · 0 + 0.5 · 0 + 0.1 · 1 = 1.1.

[0210] Calculation: Restore L6, L7:

[0211] r_L6 = 0.5 · min(0.28, 0.9) = 0.14, r_L7 = 0.14.

[0212] = 0.14 · 1 + 0.14 · 1 = 0.28.

[0213] F Calculation: F = 0.7 · 1.1 + 0.2 · 2 - 0.1 · 0.28 = 1.142.

[0214] Constraint check: Cable E2 current 2070 A > 2000 A (still violated), solution not feasible.

[0215] Solution C: Cross transformer supply, only restore L6

[0216] Operation: Open QF7 (isolate fault), close QF_Link (connect T2 system light load feeder QF9).

[0217] Calculation: = |y_QF7 - y_QF7^0| + |y_QF_Link - y_QF_Link^0| = |0 - 1| + |1 - 0| = 2.

[0218] Calculation: L5, L7, L8 de-energized (x_L5=0, x_L6=1, x_L7=0, x_L8=0):

[0219] = 1 · 1 + 0.5 · 0 + 0.5 · 1 + 0.1 · 1 = 1.6.

[0220] Calculation: Only L6 is restored, path is QF9 → QF_Link:

[0221] δ_min = min(δ_QF9, δ_QF_Link) = min(0.85, 0.9) = 0.85.

[0222] r_L6 = w_{imp,L6} · δ_min = 0.5 · 0.85 = 0.425.

[0223] = 0.425 · 1 = 0.425.

[0224] F Calculation: F = 0.7 · 1.6 + 0.2 · 2 - 0.1 · 0.425 = 1.4775.

[0225] Constraint check: QF9 current increased to 180A < 200A (all constraints satisfied), solution feasible.

[0226] Solution D: New solution, restore L6 and L7 (use fireproof cable)

[0227] Operation: Close QF_Tie, but use fireproof cable to supply power across partition (assuming available).

[0228] Calculation: = 2 (same as solution B).

[0229] Calculation: Only L5 is lost (x_L5=0, x_L6=1, x_L7=1, x_L8=0), = 1.

[0230] Calculation: Restore L6, L7, path health same as solution A, = 0.28.

[0231] F Calculation: F = 0.7 · 1 + 0.2 · 2 - 0.1 · 0.28 = 1.072.

[0232] Constraint check: If fireproof cable allows across partition, and current is not exceeded, solution is feasible (but requires additional equipment).

[0233] Feasible solution comparison: scheme C (F = 1.4775) and scheme D (F = 1.072) are feasible, but scheme D needs fireproof cable, so scheme C is selected.

[0234] Output scheduling strategy:

[0235] Fault isolation instruction: BREAKER_CONTROL (QF7, OPEN) (ensure that the fault point is isolated).

[0236] Power recovery instruction: BREAKER_CONTROL (QF_Link, CLOSE) (restore L6 through T2 system).

[0237] In an embodiment of the present application, a power supply anomaly identification and adaptive scheduling system is also provided, comprising:

[0238] A model construction module is configured to acquire a CAD power design drawing and construct a power system topology model based on a graph database.

[0239] A model correction module is configured to correct the power system topology model in real time based on sensor collected parameters to form a dynamic digital twin model.

[0240] An anomaly identification module is configured to associate an emergency lighting module and / or an image recognition module in a monitoring system, perform anomaly search and fault type identification through multi-source data fusion.

[0241] An adaptive scheduling module is configured to evaluate equipment real-time operation margin based on fault identification results and the dynamic digital twin model, and output a power supply adaptive scheduling strategy with a multi-objective optimization function.

[0242] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts of each embodiment can be referred to each other.

[0243] The skilled person can further realize that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be realized in electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been described in general terms in the above description. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0244] Finally, it should be noted that, in this document, terms such as first and second, etc., are used merely to distinguish one entity or action from another, without necessarily requiring or implying any such actual relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

Claims

1. A method for power supply anomaly identification and adaptive scheduling, characterized in that, Includes the following steps: S1: Obtain CAD power design drawings and construct a power system topology model based on a graph database. The power system topology model includes power supply paths, power switch specifications, cable diameter specifications, and sensor placement locations. S2: Correct the power system topology model based on the parameters collected by the sensor; S3: Associate the emergency lighting module and / or the image recognition module in the monitoring system, determine the abnormal retrieval area based on the abnormal signals of the data collected by the sensor, emergency lighting module and image recognition module, and perform power supply abnormality identification on the abnormal retrieval area. The identification results include power supply short circuit, overload and over-level tripping. S4: Based on the power supply anomaly identification results, evaluate the real-time operating margin and design margin of the power switching equipment. With the real-time current carrying capacity of the cable, the switch margin, electrical accessibility, spatial accessibility, and spatial fire protection zones as constraints, construct a multi-objective optimization function and output an adaptive power supply scheduling strategy.

2. The method according to claim 1, characterized in that, Step S1 specifically includes: S1.1: Analyze the CAD electrical design drawing, identify electrical component symbols and connection lines, and extract the component specifications and spatial location information; S1.2: In the graph database, electrical components are created as nodes, power supply paths are created as edges, and the specification parameters and spatial location information are used as attributes of nodes and edges; S1.3: The sensors include residual current monitoring sensors, voltage / current sensors, and / or temperature sensors, and the monitoring parameter types and placement information of the sensors are associated with the corresponding node or edge attributes.

3. The method according to claim 2, characterized in that, Step S2 specifically includes: The system collects monitoring data from various sensors in real time, associates the monitoring data with the corresponding nodes or edges in the graph database, dynamically updates the real-time status attributes of the nodes or edges, and constructs a dynamic digital twin model synchronized with the physical power grid.

4. The method according to claim 3, characterized in that, Step S3 specifically includes: S3.1: When any of the following signals are received: abnormal sensor data, activation of the emergency lighting module, or identification of abnormal visual features by the image recognition module, a preliminary abnormality retrieval area is delineated in the power system topology model; S3.2: Within the preliminary anomaly retrieval area, cross-validation is performed by querying anomaly signals from at least another data source. If the verification is successful, the preliminary anomaly retrieval area is confirmed as an anomaly retrieval area, and the fault point is accurately located using multi-source information.

5. The method according to claim 4, characterized in that, The multi-objective optimization function in step S4 is: ， in: , , These are the weighting coefficients. For load loss costs, To reduce the cost of switching operations, Benefits for power supply reliability; ； ； ； in, It is a load node in the system. For the set of load nodes, For load Importance weight, For load Power supply state variables, Power supply status. The system is in a power-off state. It's a switch In the initial state before the failure occurs It is a load Reliability score.

6. The method according to claim 5, characterized in that, The constraints in step S4 include: ； ； ； ； ； ； ； in, Power to load The set of switches on possible paths, For power supply collection, Indicates the number of power nodes. For flow through the switch Real-time current, For switch Rated current, For the first Real-time current of each device For the first The rated current carrying capacity of the power supply cable for each device. This is a temperature correction factor. Indicates switch Electrical operability, Indicates switch Physical operability, Indicates load Is it powered by a power source? powered by, Ensure power accessibility for fire-resistant zones.

7. The method according to claim 6, characterized in that, In step S4, the multi-objective optimization function is solved using a mixed-integer linear programming (MILP) solver, including Gurobi, CPLEX, or SCIP. The solver takes the objective function, constraints, and real-time parameters as input and outputs the optimal combination of switching states. The combination is then converted into a sequence of switching operation commands and sent to the power monitoring system.

8. A power supply anomaly identification and adaptive scheduling system, used to execute the power supply anomaly identification and adaptive scheduling method as described in any one of claims 1-7, characterized in that, include: The model building module is used to obtain CAD power design drawings and build a power system topology model based on a graph database. The model correction module is used to correct the power system topology model in real time based on the parameters collected by the sensors, forming a dynamic digital twin model; Anomaly detection module is used to associate emergency lighting module and / or image recognition module in monitoring system to perform anomaly retrieval and fault type identification through multi-source data fusion; The adaptive scheduling module is used to evaluate the real-time operating margin of the equipment based on the fault identification results and the dynamic digital twin model, and output the power supply adaptive scheduling strategy with a multi-objective optimization function.

9. An electronic device, characterized in that, include: One or more processors; Memory; and one or more programs stored in memory, the one or more programs including instructions for executing the power supply anomaly identification and adaptive scheduling method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, It includes one or more programs that are executed by one or more processors of an electronic device, the one or more programs including instructions for performing the power supply anomaly identification and adaptive scheduling method as described in any one of claims 1-7.

Images