Intelligent electrical fire risk assessment method based on data analysis

Abstract

The application provides an intelligent electrical fire risk assessment method based on data analysis, which is applied to the field of electrical safety and risk management, acquires electrical system structure information and multi-dimensional operation data, constructs a hierarchical power supply model, analyzes power transmission deviation of the power distribution layer and the load layer, and combines the cumulative state quantity of the power distribution layer, the connection layer structure characteristics and the load layer power consumption behavior characteristics to perform multi-level risk calculation and correction, thereby generating a comprehensive fire risk value, and realizing quantitative evaluation and early warning of potential fire hazards of the electrical system.

Description

Technical Field

[0001] This application relates to the field of electrical safety and risk management technology, and in particular to a data analysis-based intelligent electrical fire risk assessment method. Background Technology

[0002] With the continuous improvement of industrial electrification and the widespread application of smart grids and building electrical systems, electrical fires have become one of the major hidden dangers threatening production safety and the safety of people's lives and property. Electrical fires are often caused by factors such as overload, poor contact, insulation aging, and abnormal harmonics. They are characterized by strong concealment, slow evolution, and suddenness. Once they occur, the consequences are serious.

[0003] In existing electrical fire risk assessment and prevention, methods such as regular manual inspections, threshold alarm devices based on single parameters, and some power quality monitoring equipment are mainly used to complete the risk assessment and management of electrical fires.

[0004] However, in existing electrical fire risk assessment and prevention practices, hidden faults caused by factors such as the slow decline in the insulation performance of electrical circuits and the gradual degradation of the contact performance of connection parts often present as long-term, low-amplitude energy loss and thermal stress accumulation. In the early stages, their impact on conventional monitoring parameters such as current and temperature is not significant, and the changes in related parameters are usually within the preset alarm threshold range, making it difficult to trigger the alarm mechanism of the above-mentioned methods. Therefore, under the current technical conditions, the energy anomalies and risk accumulation processes caused by these hidden faults are difficult to detect in a timely manner, thus keeping the relevant fault hazards unidentified for a long time. As a result, the relevant hazards are often only identified after they have developed to a more serious stage, leading to a high rate of missed reports in electrical fire risk assessment. Summary of the Invention

[0005] This application provides a data analysis-based intelligent electrical fire risk assessment method. Its core lies in: constructing a hierarchical power supply model (distribution layer, connection layer, and load layer) by hierarchically collecting and analyzing electrical system structural information and operational data; calculating the transmission deviation of the electrical system based on power balance relationships; when the transmission deviation exceeds a threshold, weighted fusion and correction are performed by combining the cumulative state variables of the distribution layer, the risk transmission coefficient of the connection layer, and the electricity consumption behavior characteristics of the load layer to generate a comprehensive fire risk value; further, trend analysis is conducted on the changes of the comprehensive risk value over time to determine the risk evolution trend, and fire risk levels are classified accordingly to generate targeted prevention strategies, thereby achieving early identification and risk assessment of hidden faults and reducing the false negative rate of electrical fire risk assessment.

[0006] To achieve the above objectives, this application adopts the following technical solution:

[0007] This application provides a data analysis-based intelligent electrical fire risk assessment method, which may include:

[0008] Obtain the structural information and electrical data of the electrical system, and construct a hierarchical power supply model of the electrical system based on the structural information. The hierarchical power supply model includes a power distribution layer, a load layer, and a connection layer.

[0009] Based on the electrical data, the power balance relationship between the power distribution layer and the load layer is analyzed to obtain the transmission deviation of the electrical system;

[0010] If the transmission deviation is greater than a preset deviation threshold, a fire risk assessment is triggered, the operation data of the power distribution layer is obtained, the operation data is analyzed, the cumulative state quantity of the power distribution layer is calculated, and the cumulative state quantity is weighted and fused to obtain the first fire risk value of the electrical system.

[0011] Based on the structural characteristics of the connecting layer, the first fire risk value is corrected for risk transmission to obtain the second fire risk value;

[0012] The power consumption behavior data of the load layer is obtained, the power consumption behavior data is statistically analyzed to obtain the power consumption correction factor, and the second fire risk value is corrected according to the power consumption correction factor to obtain the comprehensive fire risk value.

[0013] The changes in the comprehensive fire risk value over multiple time periods are analyzed to obtain the corresponding risk evolution trend. Based on the comprehensive fire risk value and the risk evolution trend, the fire risk level is determined, and a corresponding prevention strategy is generated based on the fire risk level.

[0014] In some possible implementations, the step of analyzing the power balance relationship between the distribution layer and the load layer based on the electrical data to obtain the transmission deviation of the electrical system includes:

[0015] Based on the electrical data, the power distribution value of the power distribution layer within a preset time window is calculated;

[0016] Based on the electrical data, calculate the output power data of the load layer within a preset time window;

[0017] The power allocation value is compared with the output power data to calculate the power difference between the power distribution layer and the load layer;

[0018] The power difference is accumulated to obtain the transmission deviation of the electrical system.

[0019] In some possible implementations, the accumulation of the power difference to obtain the transmission deviation of the electrical system includes:

[0020] The power difference is recorded multiple times to form a corresponding power difference sequence;

[0021] The changes in the power difference sequence are analyzed. If the power difference sequence is monotonically increasing, a first weight is assigned to the power difference. Based on the first weight, the absolute values ​​of the power difference are weighted and accumulated to obtain the transmission deviation of the electrical system.

[0022] If the power difference sequence fluctuates around zero, a second weight is assigned to the power difference. Based on the second weight, the absolute values ​​of the power difference are weighted and accumulated to obtain the transmission deviation of the electrical system.

[0023] In some possible implementations, the operating data includes distribution current data, conductor temperature data, power quality data, and load mode data; the cumulative state quantities include current stress accumulation index, thermal accumulation index, harmonic accumulation index, and load fluctuation accumulation index; and the analysis of the operating data to calculate the cumulative state quantities of the distribution layer includes:

[0024] Based on the power distribution current data, the cumulative time when the current value is greater than the preset rated current threshold and the deviation between the current value and the preset rated current threshold are calculated to obtain the current stress accumulation index.

[0025] The temperature data of the conductor is analyzed to obtain the heat accumulation index;

[0026] Based on the power quality data, the cumulative time when the current harmonic distortion rate is greater than the preset harmonic limit and the deviation between the current harmonic distortion rate and the preset harmonic limit are calculated to obtain the harmonic accumulation index.

[0027] Based on the load pattern data, the load rate sequence of the power distribution layer is obtained, and the variance of the load rate sequence and the proportion of time in the load rate sequence where the load rate is greater than a preset load threshold are calculated to obtain the cumulative load fluctuation index.

[0028] In some possible implementations, the weighted fusion of the accumulated state quantities to obtain the first fire risk value of the electrical system includes:

[0029] The current stress accumulation index, the thermal accumulation index, the harmonic accumulation index, and the load fluctuation accumulation index are normalized and mapped to a preset normalized risk value range to obtain a set of normalized indices.

[0030] Based on preset fusion weights, the normalized index set is weighted and summed to obtain the first fire risk value of the electrical system.

[0031] In some possible implementations, the step of combining the structural features of the connecting layer to perform risk transmission correction on the first fire risk value to obtain a second fire risk value includes:

[0032] Based on the structural characteristics of the connection layer, the risk transmission coefficient between the power distribution layer and the load layer is calculated;

[0033] Based on the risk transmission coefficient, the first fire risk value is corrected to obtain the second fire risk value.

[0034] In some possible implementations, the statistical analysis of the electricity consumption behavior data to obtain an electricity consumption correction factor, and the correction of the second fire risk value based on the electricity consumption correction factor to obtain a comprehensive fire risk value, includes:

[0035] Perform time-series analysis on the electricity consumption behavior data to extract the corresponding electricity consumption characteristic quantities;

[0036] Based on the power consumption characteristics, the power consumption behavior of the load layer is analyzed to determine the corresponding power consumption status bonus value;

[0037] The electricity consumption status additive value is normalized and mapped to form an electricity consumption correction factor;

[0038] Based on the electricity consumption correction factor, the second fire risk value is amplified or suppressed to obtain a comprehensive fire risk value.

[0039] In some possible implementations, the electricity consumption characteristics include load characteristics, power characteristics, and behavioral characteristics. The step of analyzing the electricity consumption behavior of the load layer based on these characteristics to determine the corresponding electricity consumption status bonus includes:

[0040] The load level of the load layer is analyzed based on the load characteristics to obtain the first power consumption state;

[0041] The power fluctuation of the load layer is analyzed based on the power characteristic quantity to obtain the second power consumption state;

[0042] The power consumption pattern of the load layer is analyzed based on the behavioral characteristic quantities to obtain the third power consumption state;

[0043] The power consumption state type of the load layer is determined based on the first power consumption state, the second power consumption state, and the third power consumption state.

[0044] Based on the power consumption status type, determine the corresponding power consumption status bonus value.

[0045] In some possible implementations, the analysis of the changes in the comprehensive fire risk value over multiple time periods to obtain the corresponding risk evolution trend includes:

[0046] The comprehensive fire risk value is recorded over multiple time periods to form a risk time series;

[0047] Trend fitting is performed on the risk time series to obtain a risk trend line;

[0048] Calculate the slope of the risk trend line over multiple time periods to form the corresponding risk change rate;

[0049] Based on the rate of risk change, a corresponding risk evolution trend is generated.

[0050] In some possible implementations, determining the fire risk level based on the comprehensive fire risk value and the risk evolution trend, and generating a corresponding prevention strategy based on the fire risk level, includes:

[0051] If the comprehensive fire risk value is less than the first threshold, it is determined to be the first fire risk level;

[0052] If the comprehensive fire risk value is between the first threshold and the second threshold and the rate of risk change is greater than zero, it is determined to be the second fire risk level.

[0053] If the comprehensive fire risk value is between the second threshold and the third threshold, and the rate of risk change is greater than the first rate threshold, it is determined to be a third fire risk level;

[0054] If the comprehensive fire risk value is greater than the third threshold, or if the comprehensive fire risk value is greater than the fourth threshold and the rate of risk change is greater than the second rate threshold, it is determined to be the fourth fire risk level.

[0055] Based on the first fire risk level, the second fire risk level, the third fire risk level, and the fourth fire risk level, corresponding prevention strategies are generated respectively.

[0056] As can be seen from the above technical solution, this application has the following beneficial effects:

[0057] 1. This application constructs a hierarchical power supply model for the electrical system and calculates the first fire risk value based on the cumulative state variables of the distribution layer, thereby achieving a quantitative assessment of system operation anomalies. By introducing the line structure characteristics of the connection layer and using the risk transmission coefficient to spatially correct the first fire risk value, a second fire risk value is formed, thus characterizing the transmission path of fire risk in the electrical system and improving the accuracy of risk assessment.

[0058] 2. This application collects power consumption behavior data of the load layer, extracts load characteristic quantities, power characteristic quantities and behavioral characteristic quantities, comprehensively determines the power consumption status type of the load layer, generates an additive value based on the power consumption status, and forms a power consumption correction factor through normalization mapping, which amplifies or suppresses the second fire risk value, thereby realizing the adjustment of risk at the behavioral level, and thus accurately reflecting the impact of load operation habits on fire risk.

[0059] 3. This application calculates the rate of risk change and forms a risk evolution trend by performing time series analysis on the comprehensive fire risk value over multiple time periods. Based on the risk evolution trend and the comprehensive risk value, the fire risk level is determined and a corresponding prevention strategy is generated. This enables the management of risk from quantitative assessment to level determination and then to prevention and control decision-making, thereby assessing the fire risk of electrical systems and guiding actual prevention and control measures. Attached Figure Description

[0060] The present application will be further described below with reference to the accompanying drawings.

[0061] Figure 1 A flowchart of the first data analysis-based intelligent electrical fire risk assessment method provided in this application;

[0062] Figure 2 A flowchart of the second data analysis-based intelligent electrical fire risk assessment method provided in this application;

[0063] Figure 3 An example diagram illustrating the data analysis-based intelligent electrical fire risk assessment method provided in this application. Detailed Implementation

[0064] The terms "first," "second," and "third," etc., used in this application specification, claims, and drawings are used to distinguish different objects, not to limit a specific order.

[0065] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0066] To ensure clarity and conciseness in the description of the following embodiments, a brief introduction to the related technologies is given first:

[0067] Current assessments of electrical fire risks primarily rely on the following specific technical approaches:

[0068] Regular manual inspections are a commonly used basic method. Inspectors use infrared thermometers to randomly measure the temperature of key components such as distribution cabinets, cable joints, and switch contacts; they use clamp meters to measure line current to determine if there is an overload; and they visually inspect electrical equipment for abnormal conditions such as dust accumulation, corrosion, and unusual odors, manually recording the inspection results. This method relies heavily on the experience and sense of responsibility of the personnel, has long inspection intervals, limited coverage, and cannot achieve continuous detection of potential hazards.

[0069] Technically, this primarily relies on automatic alarm devices based on fixed thresholds. These systems typically involve installing residual current-based electrical fire monitoring detectors and temperature-measuring electrical fire monitoring detectors in the power distribution circuit. The former monitors the residual current (leakage current) in the circuit in real time using a residual current transformer, while the latter monitors the temperature of cables or connection points using contact or non-contact sensors. These detectors transmit the collected analog signals to a central control unit via wired (e.g., bus-based) or wireless methods. The core of the control unit is a microcontroller (such as a single-chip microcomputer or PLC) with preset fixed alarm thresholds. Its workflow involves directly comparing the real-time data from the sensors with the preset thresholds in memory. Only when the real-time data exceeds the fixed threshold will the control unit activate an audible and visual alarm or cut off the relevant power supply via a relay output control signal. The core logic of this technology is immediate alarm for exceeding the threshold; its detection and judgment are independent and point-based.

[0070] In situations with high power quality requirements, independent power quality measurement equipment is deployed. This type of equipment (such as portable or online power quality analyzers) uses high-precision sampling to measure and record parameters such as voltage and current harmonic content, three-phase imbalance, and voltage sag. It typically works by continuously recording these parameters and generating reports, which are then periodically analyzed by professionals to determine if the data exceeds national standard limits, thus assessing the additional heating risks caused by harmonics. However, this method is often independent of fire alarm systems, and the analysis is delayed, lacking real-time early warning capabilities.

[0071] Research has found that while regular manual inspections cover critical electrical equipment, they are limited by inspection frequency and personnel capabilities, making continuous monitoring of the electrical system impossible. Inspection intervals are typically on a daily or weekly basis, failing to capture subtle changes in hidden faults over a short period. Furthermore, the limited inspection coverage means some lines and connection points are easily overlooked, leading to the accumulation of hidden dangers without timely detection.

[0072] Fixed-threshold alarm devices can only identify instantaneous exceedances. Taking residual current or temperature detection as an example, these devices rely on preset thresholds to trigger alarms, with the core logic being point-based, immediate exceedance judgment. For energy leakage caused by slow degradation of electrical circuit insulation or poor high-resistance contact, the resulting current or temperature changes are small and accumulate slowly, often remaining within the threshold range for a long time without triggering an alarm, thus leading to early hazard failure. Furthermore, alarm devices typically detect single circuits or single nodes, lacking system-level analysis capabilities and failing to capture the transmission effect of localized, hidden faults on the overall power supply system risk.

[0073] While power quality measurement equipment can accurately collect and analyze parameters such as harmonics and voltage imbalances, its analysis largely relies on post-event reports or manual interpretation, lacking real-time capabilities and failing to promptly reflect fault evolution trends and potential risk accumulation. Furthermore, because these detection methods are often deployed in isolation, the lack of data fusion makes it difficult to comprehensively grasp the overall risk situation of the electrical system.

[0074] The above analysis reveals a lack of ability to identify long-term, low-amplitude, cumulative, and latent faults. Therefore, while existing methods can detect obvious anomalies or momentary exceedances, they still suffer from high rates of underreporting of early, latent faults, failing to meet the demands of modern electrical system safety management for proactive prevention and real-time assessment.

[0075] Example 1

[0076] To address the aforementioned issues, this application provides a data analysis-based intelligent electrical fire risk assessment method. Please refer to [link / reference]. Figure 1 and Figure 3 .

[0077] S101: Obtain the structural information and electrical data of the electrical system, and construct a hierarchical power supply model of the electrical system based on the structural information.

[0078] The process involves collecting structural information and electrical data from the electrical system to be evaluated. Structural information includes the distribution cabinets, busbars, circuit breakers, conductors, cables, load devices, and the connections and hierarchical relationships between these devices. Electrical data includes real-time operating parameters such as current, voltage, power, temperature, harmonics, and load patterns. Based on this structural information, the electrical system is divided into multiple layers, and a hierarchical power supply model is constructed. This model includes a distribution layer, a connection layer, and a load layer, which are used for subsequent power transfer analysis and fire risk assessment.

[0079] Among them, the hierarchical power supply model refers to a multi-level power supply representation model established based on electrical system structure information, which is used to simulate the power flow and risk transmission path of power from the distribution layer to the load layer.

[0080] The power distribution layer refers to the system level in an electrical system responsible for power input, distribution, and the location of major protection devices, including the main distribution cabinet, branch circuit breakers, main busbar system, and branch power feeders.

[0081] The connection layer refers to the intermediate connection medium between each power distribution unit and the load equipment in an electrical system, including terminals, terminal blocks, connectors, wires, cables, interfaces and related fasteners.

[0082] The load layer refers to the terminal equipment layer in an electrical system that ultimately consumes electrical energy, including various electrical devices such as motor-driven loads, lighting loads, control equipment, and information system equipment.

[0083] In some possible implementation methods, the process of constructing a hierarchical power supply model based on the structural information of the electrical system involves comprehensively collecting structural information about the electrical system through design drawings, building information models (BIMs), electrical wiring diagrams, and equipment manuals. Specifically, design drawings provide the overall topology layout of the electrical system, including the location and connection relationships of distribution cabinets; BIMs are used to extract three-dimensional spatial coordinates and building environment factors, such as floor distribution and fire compartments; electrical wiring diagrams record in detail the wiring paths, switch configurations, and cable specifications; and equipment manuals supplement node attributes, such as transformer capacity, rated voltage, material type, and aging factor. The information collected from these sources is digitized, for example, by using software tools (such as CAD or BIM software) to scan and convert drawings into vector format, or by exporting data from databases, ensuring information integrity and consistency.

[0084] After obtaining the structural information of the electrical system, the electrical system is divided into three levels: the power distribution layer, the connection layer, and the load layer.

[0085] For example, the construction of the power distribution layer involves selecting all devices with power supply attributes from the structural information as power distribution nodes, such as main transformers, generator sets, high and low voltage switchgear, busbars, etc. For each power distribution node, the following key attributes are assigned: capacity (e.g., 500kVA), rated voltage (e.g., 10kV), rated current, equipment commissioning years, insulation material type, historical maintenance records, and initial aging factor (the aging factor is usually estimated based on the commissioning years and material standard lifespan; for example, the standard lifespan of transformer oil-paper insulation is about 30 to 40 years, and the initial value can be calculated by linearly increasing by 0.02-0.05 per year). The construction of the load layer involves selecting electrical equipment or units from the structural information as load nodes, such as motors, lighting fixtures, air conditioning units, charging piles, production equipment control cabinets, etc. Each load node is assigned the following attributes: rated power (e.g., 50kW), rated voltage, equipment type, installation location coordinates, commissioning years, operating environment conditions (e.g., high temperature and high humidity workshop or clean data center), and load characteristic description (constant power type, inductive load type, etc.). The construction of the connection layer, based on the electrical wiring diagram and design drawings in the structural information, determines the connection relationship between the power distribution nodes in the power distribution layer and the load nodes in the load layer, creating a set of directed edges, such as cable connections from power distribution nodes to load nodes. It also incorporates corresponding circuit connection methods and structural features, such as series connections of cables, busbars and switches (linear paths, marked with connection methods to represent unidirectional energy flow, no branches), parallel connections (multiple redundant paths, marked with connection methods to support load sharing calculations), ring network structures (closed-loop backup paths, marked with connection methods to simulate fault detours), and radial structures (tree-like branched paths, marked with connection methods to highlight the dependence of the central node). Each edge of the connection layer is assigned extended attributes, such as length (e.g., 50m), material (e.g., copper wire), resistivity (e.g., 0.01 ohm / m), connection method label (e.g., series or ring network), and structural feature descriptions (e.g., number of branches, redundancy coefficient, or loop depth).

[0086] After constructing the distribution layer, connection layer, and load layer, these three layers and their associated attributes can be integrated to obtain a complete hierarchical power supply model. Each distribution node in the distribution layer, each load node in the load layer, and each edge in the connection layer (i.e., the connection relationship between distribution nodes and load nodes) is represented using a unified topology based on structural information. Each node and edge contains its corresponding key and extended attributes, such as node capacity, voltage, rated power, material type, installation environment, and aging coefficient; and edge length, resistivity, connection method label, and redundancy coefficient. Based on the directed edge set of the connection layer, the distribution layer and load layer are mapped through the topological relationship of the connection layer. By using distribution nodes as source points and load nodes as sink points, paths from the distribution layer to the load layer are established along the directions of the connection layer edges, preserving the topological characteristics and structural features of each path. The entire hierarchical power supply model is then encapsulated using a data structure. Node and edge attributes can be stored using a unified data structure, such as an adjacency list, graph structure, or matrix representation, binding the attributes and inter-layer relationships of the power distribution layer, connection layer, and load layer into the same model. Through the above steps, the resulting hierarchical power supply model is a structured graph model containing node attributes, edge attributes, and inter-layer energy flow relationships. This model can systematically represent the hierarchical topology and operating characteristics of an electrical system.

[0087] S102, Based on electrical data, analyze the power balance relationship between the distribution layer and the load layer to obtain the transmission deviation of the electrical system.

[0088] Specifically, real-time electrical data from the distribution layer and load layer is acquired, including operating parameters such as current, voltage, power, and load mode. Within a preset time window, the power allocation value of each distribution node in the distribution layer is calculated, i.e., the theoretical power output from the distribution layer; simultaneously, the output power data of each load node in the load layer is calculated, i.e., the actual power consumption value of the load layer. The power allocation value of the distribution layer is compared with the actual power of the load layer to obtain the power difference between the distribution and load paths. These power differences are further accumulated to form a transmission deviation index.

[0089] Among them, the power allocation value refers to the power that the distribution layer node should theoretically output to the downstream load according to the design capacity and load allocation ratio within a specific time window.

[0090] Output power data refers to the actual electrical power consumed by each load node in the load layer within the same time window.

[0091] Power difference refers to the difference between the actual power of the distribution layer and the load layer.

[0092] Transmission deviation is an index obtained by accumulating and weighting the power difference sequence. It is used to quantify the degree of power transmission abnormality from power distribution to load and can reflect problems such as hidden energy loss or poor contact in the system.

[0093] In some possible implementation methods, real-time electrical data of the distribution layer and load layer are acquired, including current, voltage, power, power factor, load mode, and other operating parameters. For a preset time window, the power allocation value of each distribution node in the distribution layer is calculated, that is, the power that should theoretically be output according to node capacity, load allocation ratio, and topology path.

[0094] For example, the power allocation value is achieved through the following means: Capacity information of each distribution node in the distribution layer is collected, including rated power, rated current, years of operation, and historical load records. Simultaneously, the rated power, load type, and real-time load mode of downstream load nodes are obtained. Based on the topology of the connection layer, distribution nodes are used as sources and load nodes as sinks. Power transmission paths are established along each path, and the load allocation ratio is determined according to the connection method label (e.g., series, parallel, ring network, or radial). For series paths, downstream loads are allocated power according to line impedance and rated capacity ratio. For parallel paths, the allocation ratio is determined by combining load distribution characteristics and line impedance. In ring networks or redundant paths, power allocation is adjusted according to the design redundancy coefficient and load sharing strategy. The theoretical power value on each path, combined with the distribution node capacity and allocation ratio, is multiplied by the line efficiency coefficient (considering conductor impedance, voltage drop, and line loss) to obtain the preliminary power allocation value. Based on this, the theoretical power value is further time-weighted and adjusted according to load changes within a preset time window to obtain the power allocation value that each distribution node should output to the downstream load node within that time window. This power allocation value reflects the theoretical distribution of electrical energy transferred from the distribution layer to the load layer under ideal conditions.

[0095] Simultaneously, the output power data of each load node in the load layer within the same time window is calculated, i.e., the actual power consumed by the load. This is achieved by comparing the power allocation value of the distribution layer with the actual power data of the load layer (by calculating the power difference between each distribution node and its connected multiple load nodes, summing the absolute values, and obtaining the power difference between the corresponding distribution layer and load layer). The calculation method is that the power deviation is equal to the difference between the power allocation value of the distribution node and its connected multiple load nodes, and the absolute values ​​of the multiple power differences are summed to obtain the power difference between the distribution layer and the load layer. This power difference can reflect the energy deviation in the process of power distribution to the load.

[0096] In some possible implementations, the power difference calculated between the distribution layer and the load layer within a continuous time window is recorded multiple times in chronological order to form a power difference sequence, with each item corresponding to the power deviation of a time window. To quantify the degree of transmission anomaly from distribution to load, the power difference sequence needs to be weighted and accumulated. Specifically, the trend of the sequence can be analyzed first: when the power difference sequence shows a monotonically increasing trend, it indicates that the energy deviation transferred from the distribution layer to the load layer is gradually widening. At this time, each power difference in the sequence is given a higher weight (first weight, 1.2~1.5) to enhance the sensitivity to continuous accumulation anomalies. When the power difference sequence fluctuates around zero, it indicates that the power deviation has short-term fluctuations but no obvious accumulation trend. At this time, a lower weight is given (second weight, 0.6~0.9) to reduce the impact of occasional fluctuations. When the power difference sequence remains stable within a preset time window, that is, when the power difference neither shows a monotonically increasing trend nor fluctuates frequently near zero, the corresponding power difference is given a weight of 1 for direct accumulation calculation of the transmission deviation. It should be noted that the first weight and the second weight mentioned above can be set by those skilled in the art according to specific circumstances, and are not specifically limited here.

[0097] To avoid offsetting positive and negative values, the absolute value of the power difference at each time point is taken, and then multiplied by the corresponding weight to obtain the weighted power difference. By accumulating and summing the weighted power differences over the entire time window, the transmission deviation index of the electrical system can be obtained. This accumulation method not only considers the magnitude of the power deviation but also incorporates sequence trend information, making it more sensitive to continuous anomalies while suppressing occasional fluctuations or small oscillations. This allows for a more accurate reflection of potential hidden energy losses, poor contact, or line fault risks during power transfer from the distribution layer to the load layer. Alternatively, a sliding time window approach can be used to accumulate the power difference sequence in segments. Each sliding window calculates a local transmission deviation, which is then weighted and summed globally to form the final transmission deviation of the electrical system. This method can account for both short-term fluctuations and long-term accumulated anomalies, providing a reliable quantitative basis for subsequent fire risk assessment.

[0098] This application quantifies anomalies in power transmission within an electrical system by analyzing the power balance relationship between the distribution layer and the load layer. Specifically, it acquires real-time electrical data from both the distribution layer and the load layer. Within a preset time window, it calculates the theoretical power distribution value for each node in the distribution layer and the actual power consumption value for each node in the load layer. By comparing the two, it obtains the power difference, which is used to quantify the energy deviation during power distribution to the load, thereby reflecting line losses, poor contact, or hidden load anomalies.

[0099] The power difference sequence is recorded chronologically, and its trend is analyzed. A monotonically increasing power difference sequence indicates a continuously widening energy deviation, and is assigned a higher weight. A sequence fluctuating around zero is assigned a lower weight, and a stable sequence with no clear trend has a weight of 1. The absolute value of the power difference at each time point is multiplied by its corresponding weight, and then summed over the entire time window to form the transmission deviation index of the electrical system. This method combines power magnitude and trend information, making it more sensitive to continuous anomalies while suppressing interference from occasional fluctuations or small oscillations.

[0100] Through the above processing, this application can accurately reflect abnormal situations in the power transfer process from the distribution layer to the load layer, providing a reliable quantitative basis for fire risk assessment. In complex multi-path load networks, this transfer deviation index can reveal potential hidden energy losses, poor contact, or line fault risks, thereby enabling early warning and scientific prevention and control of the electrical system's operating status.

[0101] S103, if the transmission deviation is greater than the preset deviation threshold, fire risk assessment is triggered, the operation data of the power distribution layer is obtained, the operation data is analyzed, the cumulative state quantity of the power distribution layer is calculated, the cumulative state quantity is weighted and fused to obtain the first fire risk value of the electrical system.

[0102] Specifically, operational data of the power distribution layer is acquired, including distribution current data, conductor temperature data, power quality data, and load mode data. Within a preset time window, the operational data is analyzed to calculate the cumulative state variables of the power distribution layer, including cumulative current stress, cumulative thermal load, cumulative harmonic load, and cumulative load fluctuation. Each cumulative state variable is normalized, and the normalized variables are weighted and summed based on preset fusion weights to obtain the first fire risk value of the electrical system.

[0103] Among them, the operational data refers to the information such as current, voltage, conductor temperature, power quality (such as harmonic distortion rate), and load mode generated by each node in the distribution layer during real-time operation. Conductor temperature refers to the actual temperature of the conductors, buses, or equipment surface or interior of the distribution node. Power quality (such as harmonic distortion rate) refers to the degree to which the voltage and current waveforms of the distribution node deviate from the ideal sine wave, including parameters such as total harmonic distortion (THD). Load mode refers to the operating characteristics of the downstream load of the distribution node, including load type (constant power type, inductive load type, variable load type, etc.), load change frequency, and load proportion. The load mode reflects the actual power consumption changes of the distribution system.

[0104] Cumulative state quantities refer to quantitative indicators obtained through statistical processing of operational data. They are used to reflect the long-term operational stress and potential risks of electrical equipment. These include: current stress cumulative indicators, which indicate the duration and amplitude deviation of current exceeding the limit at distribution nodes; thermal cumulative indicators, which indicate the long-term thermal stress accumulation of conductors and equipment due to temperature rise; harmonic cumulative indicators, which indicate the time and amplitude deviation of current harmonic distortion exceeding the limit, reflecting the stress of abnormal power quality on equipment; and load fluctuation cumulative indicators, which indicate the degree of frequent load changes or overload operation, including load rate variance and the proportion of time exceeding the threshold.

[0105] The first fire risk value is an index obtained by weighted fusion of various cumulative state quantities, used to quantify the potential fire risk level of the power distribution layer.

[0106] Distribution current data refers to the actual current flowing through each node of the distribution layer during operation, including instantaneous current value, average current, and maximum current.

[0107] Conductor temperature data refers to the actual temperature information of the surface and interior of conductors, busbars, or equipment in power distribution nodes.

[0108] Power quality data refers to the deviation of voltage and current waveforms at distribution nodes, including parameters such as harmonic distortion rate (e.g., total harmonic distortion, THD), frequency fluctuations, and voltage sags or swells.

[0109] Load pattern data refers to the characteristics of downstream loads at a power distribution node during operation, including load type (constant power, inductive load, variable load, etc.), power change frequency, percentage, and load rate sequence.

[0110] In some possible implementations, a fire risk assessment is triggered when the transmission deviation exceeds a preset deviation threshold. For example, the preset deviation threshold is a reference value used to determine whether the power transmission from the power distribution layer to the load layer is abnormal. That is, when the transmission deviation between the power distribution layer and the load layer exceeds this threshold, a fire risk assessment is triggered.

[0111] The setting of preset deviation thresholds comprehensively considers factors such as the electrical system's design capacity, load characteristics, line losses, and safety margins. Specifically, the normal fluctuation range of transmission deviations can be statistically analyzed based on the electrical system's rated power, rated current, and historical operating data. The mean and standard deviation of the transmission deviation sequence can be calculated, and a threshold can be set using an empirical coefficient (e.g., 1.5 to 2 times). The threshold equals the sum of the products of the mean and standard deviation and the empirical coefficient, thus ensuring that fire assessments are not falsely triggered within the normal fluctuation range, while enabling timely response in cases of abnormal accumulation or excessive deviation. During the initial operation phase of the electrical system, historical load data can be used for estimation and iterative optimization of threshold settings to account for the impact of different seasons, different load modes, and equipment aging on power transmission. For multi-path or redundant network structures, local deviation thresholds can be set for each path, and then a global deviation threshold can be generated through a weighted method to adapt to differences in electrical system topology and varying local load concentrations. This threshold setting method can scientifically quantify the boundary between normal power transmission fluctuations and abnormal deviations, providing reliable triggering conditions for subsequent fire risk assessments, while taking into account both safety and system stability. It should be noted that the above-mentioned preset deviation threshold can be set by those skilled in the art according to specific circumstances, and no specific limitations are made here.

[0112] In some possible implementation methods, operational data from the power distribution layer is analyzed to calculate cumulative state quantity indicators. Specifically, for power distribution current data, the current value at each time point within a preset time window is traversed, the time periods exceeding the preset rated current threshold are accumulated, and the deviation value (the difference between the current value and the preset rated current threshold) at each excess point is calculated. The deviation values ​​at each time point are then weighted and summed or integrated to obtain the current stress accumulation index. For conductor temperature data, the surface and internal temperature sequences of conductors, buses, or equipment at power distribution nodes are obtained through real-time acquisition or historical records. Statistical analysis of the temperature sequences is performed, such as calculating the cumulative duration of temperature exceeding the safe reference value and the integral accumulation of temperature rise amplitude, to form a thermal accumulation index. For power quality data, such as current harmonic distortion rate, the harmonic content at each time point is compared, the proportion of time exceeding the preset harmonic limit and the amplitude deviation are statistically analyzed, and the accumulated results are integrated or weighted and summed to obtain the harmonic accumulation index. For load pattern data, the power consumption of downstream loads of distribution nodes within a preset time window is recorded as a load rate sequence. The variance of this sequence is calculated to reflect the frequency of load changes. At the same time, the proportion of time when the load rate is greater than the preset load threshold is statistically analyzed to obtain a cumulative load fluctuation index.

[0113] For example, the cumulative state variables can be calculated using mathematical methods such as cumulative integration, weighted summation, sliding window analysis, or time series statistical methods. They can also be corrected by combining historical equipment data, environmental factors, and node topology information. For instance, the cumulative current stress index can be normalized for nodes with different rated capacities to ensure comparability between different nodes; the cumulative thermal index can be corrected by combining the thermal properties of the conductor material; the cumulative harmonic index can be adjusted by combining the harmonic filtering or shunt characteristics of the node; and the cumulative load fluctuation index can be calculated differently according to the load type.

[0114] In some possible implementation methods, the cumulative state parameters of the power distribution layer are processed to obtain the first fire risk value of the electrical system. The cumulative current stress, thermal, harmonic, and load fluctuation parameters are normalized separately, mapping each parameter to a preset risk value range (e.g., 0~1 or 0~100) to eliminate differences in the dimensions and amplitude ranges of different parameters, making them comparable on the same scale. Normalization can employ linear normalization, i.e., subtracting the minimum value from each parameter and dividing by the difference between the maximum and minimum values; or nonlinear normalization, such as amplifying the high-risk range, to enhance sensitivity to abnormal states.

[0115] The normalized set of indicators is weighted and fused to comprehensively reflect the fire risk of power distribution nodes under different operating conditions. The fusion weights can be allocated according to the contribution of the indicators to the fire risk and the actual characteristics of the electrical system. For example, the current stress accumulation index and the heat accumulation index directly reflect the overload and overheating state of conductors and equipment, and can be assigned higher weights (e.g., 0.3~0.4); the harmonic accumulation index reflects the long-term stress of abnormal power quality on equipment, and can be assigned medium weights (e.g., 0.2~0.3); the load fluctuation accumulation index reflects the risk of load changes and instantaneous overload, and can be assigned moderate or slightly lower weights (e.g., 0.15~0.25) to ensure that the contribution of each indicator to the first fire risk value is consistent with its potential hazard level.

[0116] The first fire risk value of the electrical system is obtained by multiplying each normalized index by its corresponding weight and then summing the results. For example, the first fire risk value can be expressed as the sum of each normalized index multiplied by its weight. This method can quantify the comprehensive impact of different operating states on fire risk. In multi-node or multi-distribution unit systems, the local first fire risk value can be calculated at each node first. Then, the risk values ​​of each node can be weighted and merged according to node importance, power distribution, and topological location to generate the first fire risk value of the global electrical system, thus taking into account both local anomalies and overall system risk.

[0117] This application utilizes multi-dimensional operational data such as current, conductor temperature, power quality, and load patterns to calculate cumulative state variables including current stress, thermal stress, harmonic stress, and load fluctuations. Through normalization and weighted fusion, a single primary fire risk value is formed. This quantifies the comprehensive contribution of different operating conditions to fire risk, enabling early identification of abnormal loads or overload conditions.

[0118] In the scenario described in this application, this method can quickly trigger risk assessments for distribution nodes with abnormal transmission deviations. It reflects the combined impact of long-term operational stress and instantaneous anomalies through cumulative indicators, avoiding the limitation of single indicators failing to capture potential risks. By combining normalization and weighted fusion mechanisms, it highlights the impact of high-risk factors on the overall fire risk of the system, while also taking into account the topological differences of multi-node, multi-path electrical systems, achieving simultaneous quantification of local anomalies and overall risks. This approach provides reliable risk assessments in this scenario, offering a scientific basis for fire early warning and prevention decisions, and enhancing system safety and preventative management capabilities.

[0119] S104. Based on the structural characteristics of the connecting layer, the first fire risk value is corrected for risk transmission to obtain the second fire risk value.

[0120] In some possible implementation methods, the first fire risk value calculated from the distribution layer is corrected for risk transmission by combining the structural characteristics of the connection layer, thereby obtaining the second fire risk value. The structural characteristics of the connection layer refer to the actual line structure between the distribution layer and the load layer, that is, the extended attributes of the connection layer in S101 include line length, conductor cross-sectional area, number of branches, line impedance, redundant paths, and connection method labels, etc.

[0121] Specifically, based on the connection characteristics between the distribution layer and the load layer, the transmission capacity of each line or path, i.e., the risk transmission coefficient, is determined. The risk transmission coefficient can be set by directly considering the line length, conductor cross-sectional area, and branching situation: the shorter the line, the larger the cross-section, and the fewer the redundant paths, the higher its transmission coefficient, indicating that the abnormal state of the distribution layer is more easily transmitted to the load layer; the longer the line, the smaller the cross-section, and the more branches or redundant paths, the lower its transmission coefficient, indicating a weaker risk transmission capacity.

[0122] The second fire risk value is obtained by directly multiplying the first fire risk value by the risk transmission coefficient. For example, multiplying the first fire risk value by the risk transmission coefficient of the corresponding line yields the corrected risk value after transmission along that line to the load layer. In the case of multiple connected lines, the corrected risk values ​​corresponding to each line can be summed to obtain a comprehensive second fire risk value, reflecting the actual transmission effect of the risk from the distribution layer to the load layer.

[0123] For example, consider a node with a first fire risk value of 0.7. This node is connected to load units via a short-distance, large-section main line and a long-distance, small-section branch line. Based on the line characteristics, risk transmission coefficients are set to 0.8 and 0.3 respectively. The corrected risk value transmitted along the main line is then 0.56, and the corrected risk value transmitted along the branch line is 0.21. The corrected risk values ​​of the two lines can be summed (e.g., proportional to load capacity) to obtain the final second fire risk value, comprehensively reflecting the potential impact of the node's fire risk on downstream loads.

[0124] This application utilizes the structural information of the connection layer to quantify and transfer the local risks of the power distribution layer to the load layer, thereby correcting the spatial distribution of fire risks, ensuring that the risk assessment is closer to the actual operating state of the electrical system, and improving the accuracy of the assessment.

[0125] S105, acquire electricity consumption behavior data of the load layer, perform statistical analysis on the electricity consumption behavior data to obtain an electricity consumption correction factor, and correct the second fire risk value according to the electricity consumption correction factor to obtain a comprehensive fire risk value. (See also...) Figure 2 .

[0126] Specifically, after calculating the second fire risk value based on the operating status of the power distribution layer and the structural characteristics of the connection layer, the actual electricity consumption behavior characteristics of the load layer are introduced to correct the risk results at the behavioral level. By acquiring the electricity consumption behavior data of the load layer, time-series statistical analysis is performed on the electricity consumption behavior data to extract electricity consumption characteristic quantities reflecting the load operating status and electricity consumption habits; based on the electricity consumption characteristic quantities, the electricity consumption behavior of the load layer is judged to determine the corresponding electricity consumption status additive value; the electricity consumption status additive value is normalized and mapped to form an electricity consumption correction factor; and based on the electricity consumption correction factor, the second fire risk value is amplified or suppressed to obtain the comprehensive fire risk value.

[0127] To ensure clarity and conciseness in the description of the following embodiments, a brief introduction to the relevant terms is given first:

[0128] Electricity consumption behavior data refers to the data set generated by each power consumption unit or node in the load layer during actual operation, reflecting electricity consumption patterns and habits. This includes, but is not limited to, information such as load start-up and shutdown times, electricity consumption period distribution, power change processes, continuous load operation duration, and records of sudden electricity consumption changes. This type of data is used to characterize the operating patterns of the load layer in actual usage scenarios, rather than single instantaneous electrical parameters.

[0129] Electricity consumption characteristics refer to quantitative features extracted through statistical and time-series analysis of electricity consumption behavior data. These features characterize the electricity consumption status and risk characteristics of the load layer. Electricity consumption characteristics include three categories: load characteristics, power characteristics, and behavioral characteristics. Load characteristics reflect load scale and load level, such as average load factor, load duration, and high load percentage. Power characteristics reflect power variation characteristics, such as power fluctuation amplitude, power change frequency, and number of instantaneous power transitions. Behavioral characteristics reflect electricity consumption patterns and habits, such as the proportion of nighttime electricity consumption, long-term continuous operation characteristics, and the proportion of unplanned electricity consumption.

[0130] The power consumption status additive value refers to the risk adjustment amount pre-set for different power consumption status types based on the analysis results of power consumption behavior at the load level. It is used to quantify the amplification or suppression effect of power consumption behavior on fire risk.

[0131] The electricity consumption correction factor is a dimensionless coefficient formed by normalizing and mapping the electricity consumption status additive value, which is used to proportionally correct the second fire risk value.

[0132] In some possible implementation methods, within a preset statistical period, electricity consumption behavior data corresponding to each power consumption unit in the load layer is collected and processed. The statistical period can be on the minute, hour, or day level to cover typical electricity consumption changes in the load layer. Time-series statistical analysis is performed on the electricity consumption behavior data to extract electricity consumption characteristic quantities that can reflect the load operating status and electricity consumption habits.

[0133] For example, load characteristic quantities are extracted, load start-up and shutdown time series are statistically analyzed to calculate the number of load start-ups and shutdowns per unit time; the continuous operating time of the load is statistically analyzed to obtain the continuous operating time distribution; and based on the rated load parameters, the ratio of the actual operating power of the load to the rated power is calculated to form the average load rate and the proportion of high-load operating periods. By accumulating and statistically analyzing the load level of the load layer within the statistical period, it is determined whether the load is in a long-term high-load operating state. When the average load rate is high or the proportion of high-load operating time continues to increase, it indicates that the load layer is creating continuous current stress on the power distribution system, thereby further increasing the possibility of fire on the basis of existing operational and structural risks.

[0134] The process involves extracting power characteristics, performing sliding time window statistics on power changes, calculating the absolute difference in power changes between adjacent time periods to form the power fluctuation amplitude, counting the number of times power changes exceed a preset fluctuation threshold to form the power change frequency, and counting power abrupt events to form the instantaneous power jump count. By statistically analyzing the continuity and volatility of the power change process, it is determined whether the load operation exhibits frequent power jumps or unstable fluctuations. Frequent power fluctuations imply repeated current surges and thermal stresses, which can easily exacerbate the aging and poor contact risks of connection parts, thus representing a risk amplification factor at the level of electricity consumption behavior.

[0135] The extraction of behavioral characteristics is based on the distribution of electricity consumption during different time periods. This includes statistically analyzing the proportion of electricity consumption at night, during non-working hours, or during periods of low human activity; identifying situations where loads operate continuously for extended periods without significant load fluctuations to form continuous operation characteristic indicators; and statistically analyzing unplanned start-ups and shutdowns, and abnormal electricity consumption times to form the proportion of abnormal electricity consumption behaviors. By statistically analyzing the electricity consumption time distribution and operating modes of the load layer, the existence of atypical electricity consumption patterns can be identified, such as high-power consumption at night, excessively long continuous operation, or unplanned start-ups and shutdowns. Such behaviors are usually outside the design operating conditions and can weaken the system's heat dissipation and the effectiveness of protection device responses, thus creating a cumulative impact on fire risk.

[0136] In some possible implementation methods, the power consumption behavior of the load layer is analyzed based on power consumption characteristics to determine the corresponding power consumption status bonus value. Specifically, based on load characteristics, the average load rate, the proportion of high-load operating periods, and the distribution of continuous operating time within the statistical period are used as the basis for load level judgment. When the average load rate is within a preset safe range, the proportion of high-load operating periods is lower than a first preset proportion threshold, and the distribution of continuous operating time shows intermittent operating characteristics, the first power consumption status is determined to be a normal load status; when the average load rate is close to the rated load limit, or the proportion of high-load operating periods exceeds the first preset proportion threshold, and the continuous operating time is significantly extended, the first power consumption status is determined to be a high-load operating status; when the load rate remains at a high level for a long period of time, and the continuous operating time exceeds the preset safe operating time, the first power consumption status is determined to be an overload risk status.

[0137] Based on power characteristic quantities, the power fluctuation amplitude, power change frequency, and instantaneous power transition count are jointly judged. When the power fluctuation amplitude is within a preset stable range, the power change frequency is low, and the number of instantaneous power transition events is small, the second power consumption state is determined to be a stable operating state. When the power fluctuation amplitude increases significantly, the power change frequency exceeds the second preset number threshold, or instantaneous power transition events occur multiple times within the statistical period, the second power consumption state is determined to be a fluctuating operating state. When the power change shows frequent and large-amplitude transitions, and power mutation events occur in a concentrated manner within a short period of time, the second power consumption state is determined to be a violently fluctuating state.

[0138] Based on behavioral characteristics, statistical judgments are made on the proportion of electricity consumption during nighttime or non-working hours, continuous operation characteristic indicators, and the proportion of abnormal electricity consumption behavior. When electricity consumption is mainly concentrated during normal working hours, the proportion of electricity consumption during nighttime or non-working hours is low, and the continuous operation time matches the load fluctuation characteristics, the third electricity consumption state is determined to be a normal electricity consumption state. When the proportion of electricity consumption during nighttime or non-working hours is significantly increased, or the load is in continuous operation for a long time but lacks corresponding power regulation characteristics, the third electricity consumption state is determined to be an atypical electricity consumption state. When there is a high proportion of unplanned start-stop behavior, abnormal electricity consumption time points, or electricity consumption behavior that does not conform to the operation plan, the third electricity consumption state is determined to be an abnormal electricity consumption state.

[0139] By comprehensively analyzing these three types of power consumption states, the overall power consumption state type of the load layer is determined, and a power consumption state additive value is generated accordingly. Specifically, the first, second, and third power consumption states are used as three independent discrimination dimensions to perform combined analysis on the power consumption behavior of the load layer. When all three power consumption states are within the normal or stable range, the power consumption state type is determined to be a low-risk power consumption state; when at least one power consumption state exhibits a high-load operation state, a fluctuating operation state, or an atypical power consumption state, the power consumption state type is determined to be a medium-risk power consumption state; when multiple power consumption states simultaneously exhibit an overload risk state, a drastic fluctuation state, or an abnormal power consumption state, the power consumption state type is determined to be a high-risk power consumption state.

[0140] Assign a power consumption status bonus value to each type of power consumption status to quantify its impact on fire risk. For example, a low-risk power consumption status can be assigned a lower bonus value (e.g., 0.8), a medium-risk power consumption status a medium bonus value (e.g., 1.0), and a high-risk power consumption status a higher bonus value (e.g., 1.2). These values ​​can be set based on historical operating data of the load layer, fire accident statistics, and empirical analysis to reflect the amplification or suppression effect of different power consumption behaviors on fire risk.

[0141] The added value of electricity usage status is normalized to form an electricity usage correction factor. The normalization mapping can be done by linearly mapping the added value to a certain range, such as 0.8 to 1.2. This ensures that when the correction factor is less than 1, it suppresses the second fire risk value; when it equals 1, it maintains the original risk value; and when it is greater than 1, it amplifies the risk. The principle of normalization is to ensure that the risk impact of different electricity usage status types is comparable under the same dimension, while avoiding excessive amplification or suppression. The electricity usage correction factor is applied to the second fire risk value to amplify or suppress it, thus obtaining the comprehensive fire risk value. Specifically, the second fire risk value is multiplied by the electricity usage correction factor to obtain the comprehensive risk value. For example, if the second fire risk value of an electrical system is 0.7, and the power usage status type is medium risk with a power usage correction factor of 1.0, then the overall fire risk value remains 0.7; if the power usage status type is high risk with a corresponding correction factor of 1.2, then the overall fire risk value is amplified to 0.84; if the power usage status type is low risk with a corresponding correction factor of 0.8, then the overall fire risk value is reduced to 0.56.

[0142] This application can adjust the fire risk based on the structural risks of the power distribution layer and the connection layer, and further combine the actual power consumption characteristics of the load layer. This makes the comprehensive fire risk value reflect not only the structure and operation of the electrical system, but also the risk amplification or suppression effect at the behavioral level, ensuring that the risk assessment is closer to the real operating environment.

[0143] S106 analyzes the changes in the comprehensive fire risk value over multiple time periods to obtain the corresponding risk evolution trend. Based on the comprehensive fire risk value and the risk evolution trend, the fire risk level is determined, and corresponding prevention strategies are generated based on the fire risk level.

[0144] The comprehensive fire risk value refers to the numerical value obtained by quantitatively assessing the fire risk of an electrical system after considering the operating status of the power distribution layer, the structural characteristics of the connection layer, and the electrical consumption behavior characteristics of the load layer.

[0145] Risk time series refers to the sequence data formed by recording the comprehensive fire risk value over multiple consecutive time periods; risk trend line refers to the curve obtained by fitting the risk time series; risk change rate refers to the slope of the risk trend line in the time dimension.

[0146] Fire risk level is the result of classifying the fire risk of electrical systems based on the comprehensive fire risk value and the rate of risk change.

[0147] Prevention strategies are operational and management measures developed for different fire risk levels.

[0148] In some possible implementation methods, risk time series are formed by recording risk values ​​over multiple time periods. By performing trend fitting on the risk time series, a risk trend line reflecting the direction of fire risk changes can be obtained. Trend fitting can employ linear fitting or piecewise fitting methods to smooth short-term fluctuations and highlight long-term change patterns, thereby obtaining the overall evolution characteristics of the risk. Further calculation of the slope of the risk trend line in each time period yields the rate of risk change. Based on the comprehensive fire risk value and risk evolution trend, the fire risk of the electrical system is classified and assessed, and corresponding prevention strategies are generated.

[0149] Specifically, a first threshold, a second threshold, a third threshold, and a fourth threshold can be set to correspond to the dividing points of low, medium, high, and extremely high fire risks, respectively. For example, the first threshold can be set to 0.5, the second threshold to 0.65, the third threshold to 0.8, and the fourth threshold to 0.9. The threshold values ​​can be adjusted based on the actual operating experience of the electrical system, the load scale, and historical risk data.

[0150] During the risk level determination process, if the comprehensive fire risk value is less than the first threshold, it is determined to be at the first fire risk level, indicating a low-risk state. At this time, only routine inspections and periodic patrols are required. If the comprehensive fire risk value is between the first and second thresholds, and the risk change rate is greater than zero, it is determined to be at the second fire risk level, indicating that the fire risk is in an upward phase. At this time, early warnings should be triggered, patrols should be strengthened, and key equipment should be inspected. When the comprehensive fire risk value is between the second and third thresholds, and the risk change rate is greater than the first rate threshold (e.g., 0.02 / hour), it is determined to be at the third fire risk level, indicating a high risk and a significant increase. Measures such as load regulation, isolation of critical equipment, and emergency response by operators should be initiated. When the comprehensive fire risk value is greater than the third threshold, or the comprehensive fire risk value is greater than the fourth threshold and the risk change rate is greater than the second rate threshold (e.g., 0.03 / hour), it is determined to be at the fourth fire risk level, indicating an extremely high fire risk. Emergency plans need to be implemented immediately, including power outages, isolation of critical lines, and activation of fire-fighting measures.

[0151] This application provides a data analysis-based intelligent electrical fire risk assessment method. Its core principle is to decompose the operational risks of electrical systems into three levels: the distribution layer, the connection layer, and the load layer. Through hierarchical modeling, data-driven analysis, and behavior correction, the method can achieve quantitative assessment and dynamic prediction of fire risks. In its implementation, the method acquires the structural information and real-time operating data of the electrical system, constructs a hierarchical power supply model, analyzes the power balance relationship between the distribution layer and the load layer, and triggers a risk assessment when the transmission deviation exceeds the allowable range. At the distribution layer, the accumulated risk of local equipment anomalies is quantified through cumulative state quantity calculation and weighted fusion to obtain a first fire risk value. Combining the line structure characteristics of the connection layer, the first fire risk value is spatially corrected using a risk transmission coefficient to form a second fire risk value, reflecting the actual transmission effect of distribution layer anomalies to the load layer. At the load layer, by collecting electricity consumption behavior data, extracting load characteristic quantities, power characteristic quantities, and behavioral characteristic quantities, an electricity consumption state additive value is generated and normalized to form an electricity consumption correction factor. This factor amplifies or suppresses the second fire risk value at the behavioral level, resulting in a comprehensive fire risk value. The trend of the comprehensive risk value over multiple time periods is analyzed, the rate of risk change is calculated, and the fire risk level is determined based on the risk level threshold, thereby generating corresponding prevention strategies and realizing management from risk detection to prevention and control decision-making. This application, through hierarchical modeling and multidimensional data analysis, closely links local abnormal states with overall system risks. It introduces connection layer structural features and load behavior features to achieve risk transmission correction and behavior adjustment, and realizes dynamic early warning and risk level classification through trend analysis. This enables prevention strategies to proactively match actual risk changes, thereby timely identifying high load or abnormal nighttime power consumption behavior in scenarios such as industrial plants or large building electrical systems and triggering load adjustment, critical line isolation, or emergency plan execution. This effectively reduces the probability of fire and potential losses, providing an operable and dynamically responsive fire risk management solution.

[0152] The foregoing has shown and described the basic principles, main features, and advantages of this application. Those skilled in the art should understand that this application is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of this application. Various changes and modifications can be made to this application without departing from the spirit and scope thereof, and all such changes and modifications fall within the scope of this application as claimed. The scope of protection of this application is defined by the appended claims and their equivalents.

Claims

1. A data-driven intelligent electrical fire risk assessment method, characterized in that, The method includes: Obtain the structural information and electrical data of the electrical system, and construct a hierarchical power supply model of the electrical system based on the structural information. The hierarchical power supply model includes a power distribution layer, a load layer, and a connection layer. Based on the electrical data, the power balance relationship between the power distribution layer and the load layer is analyzed to obtain the transmission deviation of the electrical system; If the transmission deviation is greater than a preset deviation threshold, a fire risk assessment is triggered, the operation data of the power distribution layer is obtained, the operation data is analyzed, the cumulative state quantity of the power distribution layer is calculated, and the cumulative state quantity is weighted and fused to obtain the first fire risk value of the electrical system. Based on the structural characteristics of the connecting layer, the first fire risk value is corrected for risk transmission to obtain the second fire risk value; The power consumption behavior data of the load layer is obtained, the power consumption behavior data is statistically analyzed to obtain the power consumption correction factor, and the second fire risk value is corrected according to the power consumption correction factor to obtain the comprehensive fire risk value. The changes in the comprehensive fire risk value over multiple time periods are analyzed to obtain the corresponding risk evolution trend. Based on the comprehensive fire risk value and the risk evolution trend, the fire risk level is determined, and a corresponding prevention strategy is generated based on the fire risk level.

2. The method according to claim 1, characterized in that, The step of analyzing the power balance relationship between the distribution layer and the load layer based on the electrical data to obtain the transmission deviation of the electrical system includes: Based on the electrical data, the power distribution value of the power distribution layer within a preset time window is calculated; Based on the electrical data, calculate the output power data of the load layer within a preset time window; The power allocation value is compared with the output power data to calculate the power difference between the power distribution layer and the load layer; The power difference is accumulated to obtain the transmission deviation of the electrical system.

3. The method according to claim 2, characterized in that, The step of accumulating the power difference to obtain the transmission deviation of the electrical system includes: The power difference is recorded multiple times to form a corresponding power difference sequence; The changes in the power difference sequence are analyzed. If the power difference sequence is monotonically increasing, a first weight is assigned to the power difference. Based on the first weight, the absolute values ​​of the power difference are weighted and accumulated to obtain the transmission deviation of the electrical system. If the power difference sequence fluctuates around zero, a second weight is assigned to the power difference. Based on the second weight, the absolute values ​​of the power difference are weighted and accumulated to obtain the transmission deviation of the electrical system.

4. The method according to claim 1, characterized in that, The operational data includes distribution current data, conductor temperature data, power quality data, and load mode data. The cumulative state quantities include current stress accumulation index, thermal accumulation index, harmonic accumulation index, and load fluctuation accumulation index. Analyzing the operational data and calculating the cumulative state quantities of the distribution layer includes: Based on the power distribution current data, the cumulative time when the current value is greater than the preset rated current threshold and the deviation between the current value and the preset rated current threshold are calculated to obtain the current stress accumulation index. The temperature data of the conductor is analyzed to obtain the heat accumulation index; Based on the power quality data, the cumulative time when the current harmonic distortion rate is greater than the preset harmonic limit and the deviation between the current harmonic distortion rate and the preset harmonic limit are calculated to obtain the harmonic accumulation index. Based on the load pattern data, the load rate sequence of the power distribution layer is obtained, and the variance of the load rate sequence and the proportion of time in the load rate sequence where the load rate is greater than a preset load threshold are calculated to obtain the cumulative load fluctuation index.

5. The method according to claim 4, characterized in that, The weighted fusion of the accumulated state quantities to obtain the first fire risk value of the electrical system includes: The current stress accumulation index, the thermal accumulation index, the harmonic accumulation index, and the load fluctuation accumulation index are normalized and mapped to a preset normalized risk value range to obtain a set of normalized indices. Based on preset fusion weights, the normalized index set is weighted and summed to obtain the first fire risk value of the electrical system.

6. The method according to claim 1, characterized in that, The method of combining the structural characteristics of the connecting layer to perform risk transmission correction on the first fire risk value to obtain a second fire risk value includes: Based on the structural characteristics of the connection layer, the risk transmission coefficient between the power distribution layer and the load layer is calculated; Based on the risk transmission coefficient, the first fire risk value is corrected to obtain the second fire risk value.

7. The method according to claim 1, characterized in that, The statistical analysis of the electricity consumption behavior data yields an electricity consumption correction factor. Based on this correction factor, the second fire risk value is corrected to obtain a comprehensive fire risk value, including: Perform time-series analysis on the electricity consumption behavior data to extract the corresponding electricity consumption characteristic quantities; Based on the power consumption characteristics, the power consumption behavior of the load layer is analyzed to determine the corresponding power consumption status bonus value; The electricity consumption status additive value is normalized and mapped to form an electricity consumption correction factor; Based on the electricity consumption correction factor, the second fire risk value is amplified or suppressed to obtain a comprehensive fire risk value.

8. The method according to claim 7, characterized in that, The electricity consumption characteristics include load characteristics, power characteristics, and behavioral characteristics. The analysis of the electricity consumption behavior of the load layer based on these characteristics to determine the corresponding electricity consumption status bonus includes: The load level of the load layer is analyzed based on the load characteristics to obtain the first power consumption state; The power fluctuation of the load layer is analyzed based on the power characteristic quantity to obtain the second power consumption state; The power consumption pattern of the load layer is analyzed based on the behavioral characteristic quantities to obtain the third power consumption state; The power consumption state type of the load layer is determined based on the first power consumption state, the second power consumption state, and the third power consumption state. Based on the power consumption status type, determine the corresponding power consumption status bonus value.

9. The method according to claim 1, characterized in that, The analysis of the changes in the comprehensive fire risk value over multiple time periods yields the corresponding risk evolution trend, including: The comprehensive fire risk value is recorded over multiple time periods to form a risk time series; Trend fitting is performed on the risk time series to obtain a risk trend line; Calculate the slope of the risk trend line over multiple time periods to form the corresponding risk change rate; Based on the rate of risk change, a corresponding risk evolution trend is generated.

10. The method according to claim 9, characterized in that, The step of determining the fire risk level based on the comprehensive fire risk value and the risk evolution trend, and generating corresponding prevention strategies based on the fire risk level, includes: If the comprehensive fire risk value is less than the first threshold, it is determined to be the first fire risk level; If the comprehensive fire risk value is between the first threshold and the second threshold and the rate of risk change is greater than zero, it is determined to be the second fire risk level. If the comprehensive fire risk value is between the second threshold and the third threshold, and the rate of risk change is greater than the first rate threshold, it is determined to be a third fire risk level; If the comprehensive fire risk value is greater than the third threshold, or if the comprehensive fire risk value is greater than the fourth threshold and the rate of risk change is greater than the second rate threshold, it is determined to be the fourth fire risk level. Based on the first fire risk level, the second fire risk level, the third fire risk level, and the fourth fire risk level, corresponding prevention strategies are generated respectively.

Images