Box-type substation operation state early warning method based on multi-modal data

Abstract

The invention belongs to the technical field of power equipment state monitoring, and discloses a box-type substation operation state early warning method based on multi-modal data. According to the method, infrared thermal imaging, load current and temperature permeation multi-mode data are firstly collected, complete box body internal space thermal field distribution is constructed through a heat conduction inversion and calibration technology, the thermal interaction flux between devices and the spontaneous heating power of each device are accurately calculated, and accurate evaluation of the comprehensive thermal load is achieved. According to the method, thermal stress accumulation rate calculation is innovatively introduced, the insulation residual life consumption rate is evaluated in combination with insulation material characteristics, key early warning target equipment is identified through electrical topology analysis, and finally scientific grading early warning signals are generated based on life estimated values and influence ranges in a two-dimensional mode. The method accurately quantifies the thermal state and life consumption process of the equipment, improves the accuracy and pertinence of early warning, effectively reduces the failure rate, prolongs the service life of the equipment, and improves the reliability of a power distribution system.

Description

Technical Field

[0001] This invention relates to the field of power equipment condition monitoring technology, and more specifically, to a method for early warning of the operating status of prefabricated substations based on multimodal data. Background Technology

[0002] Existing monitoring technologies for the operational status of prefabricated substations suffer from multiple limitations, resulting in poor early warning effects and a lack of scientific basis for operation and maintenance decisions. Current technologies generally employ distributed single-point temperature sensors or simple current monitoring, making it difficult to construct a complete thermal field distribution map inside the sealed enclosure, creating large monitoring blind spots. In actual operation, narrow spaces such as corners of the enclosure and gaps between equipment often become hotspots, but are often overlooked due to the inability to deploy sensors. For example, a 10kV prefabricated substation in an industrial area experienced a sudden low-voltage side connector meltdown under normal operating parameters; subsequent investigation revealed that the meltdown was caused by heat accumulation in a corner due to poor ventilation. Furthermore, existing technologies treat each piece of equipment as an independent entity, completely ignoring the thermal interaction effects in a tightly packed environment. In actual operation, the heat released by the transformer often causes abnormal temperature increases in adjacent switches, and single-equipment monitoring cannot detect this correlated risk. In the time dimension, traditional technologies only focus on whether instantaneous parameters exceed limits, lacking quantitative assessment of long-term thermal stress accumulation effects, leading to severe internal insulation degradation in many devices even when surface parameters appear normal. Regarding equipment lifespan prediction, existing models are oversimplified, failing to consider material properties, load history, and environmental factors. This leads to increased prediction errors, resulting in inaccurate maintenance strategies, premature intervention wasting resources, or delayed handling causing failures. In actual operation and maintenance, when multiple devices simultaneously exhibit abnormal conditions, traditional early warning systems cannot distinguish processing priorities, failing to consider the importance of the equipment's location in the power supply system and the availability of backup paths. For example, if multiple devices in a residential substation issue warnings simultaneously, maintenance personnel cannot determine which devices are most critical. As a result, while handling secondary devices, the circuit breaker on the main power supply path fails, causing a large-scale power outage. Early warning grading mechanisms are simplistic and crude, usually based on a single threshold, failing to balance the dual factors of fault urgency and impact scope. This results in a lack of scientific basis for operation and maintenance decisions and low resource allocation efficiency, which is particularly prominent during peak electricity demand periods. When multiple transformer substations require maintenance, it is impossible to determine a reasonable order of handling, increasing systemic risks.

[0003] In view of this, the present invention proposes a pre-warning method for the operation status of prefabricated substations based on multimodal data to solve the above problems. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a prefabricated substation operation status early warning method based on multimodal data, comprising: Infrared thermal imaging data and load current data of each device inside the sealed box of the prefabricated substation were collected, as well as temperature penetration data of multiple measuring points on the outer wall of the box. Based on temperature penetration data and the thermal conductivity characteristics of the enclosure wall material, the spatial thermal field distribution of each area inside the enclosure is calculated by inversion, and the spatial thermal field distribution is calibrated and corrected by combining infrared thermal imaging data. Based on the calibrated and corrected spatial thermal field distribution, the heat exchange flux between each device and its adjacent devices is calculated, and the self-heating power of each device is calculated based on the load current data of each device. The heat exchange flux and self-heating power are superimposed to obtain the comprehensive heat load value of each device, and the thermal stress accumulation rate of each device is calculated based on the time-series change of the comprehensive heat load value. The insulation remaining life consumption rate of each piece of equipment is calculated based on the thermal stress accumulation rate and the thermal degradation sensitivity coefficient of the insulation material of each piece of equipment. Based on the insulation remaining life consumption rate and the series connection position of each device in the high and low voltage side electrical circuits of the prefabricated substation, the warning target device whose insulation remaining life consumption rate exceeds the preset consumption threshold and is located on the main power supply path is identified. Based on the estimated remaining insulation life of the target equipment and the outage load capacity of the prefabricated substation caused by its failure, a graded early warning signal is generated and output to the power distribution operation and maintenance system.

[0005] The technical effects and advantages of the prefabricated substation operation status early warning method based on multimodal data of this invention are as follows: This invention eliminates blind spots in traditional monitoring, enabling maintenance personnel to clearly understand the true operating status of equipment. In-depth analysis of the thermal interaction mechanism reveals hidden thermal chain relationships between devices, allowing the system to anticipate potential heat accumulation risks and intervene in advance, rather than passively waiting for faults to occur. Precise quantification of the long-term thermal stress accumulation effect makes the insulation aging process clearly visible, providing a scientific basis for equipment lifecycle management. Accurate identification of early warning targets significantly improves the value of early warnings, eliminates frequent invalid warnings, and allows maintenance resources to be precisely allocated to truly critical equipment. A scientific early warning classification system enables decision-makers to clearly grasp the urgency and scope of impact of risks, quickly determining the optimal intervention strategy in complex and ever-changing maintenance environments. In practical applications, this invention effectively extends equipment lifespan, reduces unplanned power outages and the frequency of emergency repairs, and improves the reliability and stability of the power distribution system. Attached Figure Description

[0006] Figure 1 This is a schematic diagram of a prefabricated substation operation status early warning method based on multimodal data according to the present invention. Detailed Implementation

[0007] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0008] Please see Figure 1 In this embodiment of the invention, a method for early warning of the operating status of a prefabricated substation based on multimodal data includes: Infrared thermal imaging data and load current data of each device within the enclosed enclosure of the prefabricated substation were collected, along with temperature penetration data from multiple measuring points on the outer wall of the enclosure. The infrared thermal imaging data recorded the surface temperature distribution of the equipment, the load current data reflected the real-time operating load level of the equipment, and the temperature penetration data reflected the distribution characteristics of heat transfer across the enclosure wall. These data were acquired in real-time through a multi-channel data acquisition system, including an infrared thermal imager installed inside the enclosure, current transformers, and a uniformly distributed network of temperature sensors on the outer wall. The infrared thermal imaging equipment captured the surface temperature distribution of the visible equipment area, the current transformers collected the real-time current values ​​of each major equipment circuit, and the temperature sensor network was arranged at a certain density on the outer wall to capture the temperature distribution after heat conduction through the enclosure wall. This multimodal data provided comprehensive raw information for subsequent analysis, ensuring the integrity and accuracy of the early warning method.

[0009] Based on temperature penetration data and the thermal conductivity characteristics of the enclosure wall material, the spatial thermal field distribution of each region inside the enclosure is calculated by inversion, and then calibrated and corrected using infrared thermal imaging data. The spatial thermal field distribution reflects the three-dimensional distribution of heat inside the enclosure and is fundamental to assessing the equipment's thermal environment. The inversion calculation first uses the thermal conductivity and thickness parameters of the wall material to deduce the internal heat flux density distribution from external temperature penetration data using Fourier's law of heat conduction. Then, using the heat source superposition method and combining it with the internal geometric layout of the enclosure, a complete spatial thermal field distribution model is generated. Calibration and correction are performed by verifying and adjusting the theoretical inversion results using measured temperature values ​​provided by infrared thermal imaging data, improving the accuracy of the spatial thermal field distribution. This step provides a reliable thermal environment data foundation for subsequent thermal interaction analysis.

[0010] Based on the calibrated and corrected spatial thermal field distribution, the heat exchange flux between each device and its adjacent devices is calculated, and the self-heating power of each device is calculated based on its load current data. Heat exchange flux reflects the heat exchange between devices, while self-heating power quantifies the heat generated by the device itself. The heat exchange flux calculation is based on the temperature difference between the device surface and the surrounding space, as well as the heat transfer characteristics, considering the combined effects of convection, radiation, and conduction. The self-heating power calculation selects different heating models according to the device type, combining measured load current with device parameters to accurately estimate the heat generated by the device during operation. These two calculation results together constitute a complete analysis of the device's thermal load, providing fundamental data for subsequent thermal stress assessment.

[0011] The combined heat flux and self-heating power are superimposed to obtain the comprehensive heat load value for each device. The thermal stress accumulation rate is then calculated based on the time-series variation of the comprehensive heat load value. The comprehensive heat load value is a comprehensive indicator of the device's thermal state, while the thermal stress accumulation rate reflects the rate at which the heat load causes damage to the device. The comprehensive heat load value is obtained by subtracting the heat flux from the self-heating power; a positive value indicates heat accumulation, and a negative value indicates heat loss. The thermal stress accumulation rate calculation considers the portion of the heat load exceeding the device's heat resistance baseline, quantifying the cumulative effect of long-term thermal stress through time-series integration. These two indicators together reflect the device's thermal stress state, providing crucial input for subsequent insulation life assessment.

[0012] Based on the thermal stress accumulation rate and the thermal degradation sensitivity coefficient of the insulation materials of each piece of equipment, the remaining insulation life consumption rate of each piece of equipment is calculated. The remaining insulation life consumption rate is a core indicator for assessing the health status of equipment, directly reflecting the degree of impact of thermal stress on equipment lifespan. The calculation process is based on the heat resistance characteristics of the insulation materials of each piece of equipment. The thermal stress accumulation rate is multiplied by the thermal degradation sensitivity coefficient to obtain the insulation loss component caused by thermal factors. Simultaneously, the load factor is considered; when the load exceeds the rated value, an overload additional loss component is added. The sum of these two factors yields the final remaining insulation life consumption rate, providing a quantitative basis for early warning target identification.

[0013] Based on the remaining insulation life consumption rate and the series connection position of each device in the high- and low-voltage electrical circuits of the prefabricated substation, warning target devices with an insulation remaining life consumption rate exceeding a preset consumption threshold and located on the main power supply path are identified. Identifying warning target devices is a crucial step in accurately locating high-risk equipment, comprehensively considering the health status of the equipment and its importance to the system. The identification process first constructs an electrical main path topology map to clarify the location and importance of equipment in the power supply link; then, it filters abnormal devices with an insulation remaining life consumption rate exceeding the threshold; finally, it determines whether these abnormal devices have parallel backup paths, marking abnormal devices without backup paths as warning targets to ensure the accuracy and criticality of the warnings.

[0014] Based on the estimated remaining insulation life of the target equipment and the resulting outage load capacity of the prefabricated substation after a failure, a tiered early warning signal is generated and output to the distribution operation and maintenance system. The tiered early warning signal is the final output of the system, providing intuitive risk assessment results for operation and maintenance decisions. The early warning signal generation process first calculates the estimated remaining life of the target equipment, extrapolating the future failure time based on the current consumption rate; then, it assesses the impact of the equipment failure on the system, calculating the potential outage load capacity; finally, based on the urgency of the time and the severity of the impact, it determines the final early warning level, generating complete early warning information including risk level, estimated life, impact range, and recommended measures, providing a scientific basis for decision-making in the distribution operation and maintenance system.

[0015] In this embodiment of the invention, the detailed implementation steps for inverting and calculating the spatial thermal field distribution of each region inside the box based on temperature permeation data and the thermal conductivity characteristics of the box wall material include: The thermal conductivity and wall thickness parameters of each wall surface of the enclosure are obtained. Material thermal conductivity and wall thickness are fundamental parameters for heat conduction calculations, directly affecting the rate and distribution of heat transfer. The acquisition process first extracts basic parameters for each wall surface from the design specifications and material specifications of the prefabricated substation; then, adjustments are made based on the actual installation environment and service life, considering material aging and environmental impacts; finally, a parameter database is established, containing key information such as material type, thermal conductivity, thickness, and area for each wall region. For composite material walls, an equivalent thermal conductivity calculation method is used to simplify multi-layered materials into a single parameter representation. These parameters provide an accurate physical basis for subsequent heat conduction calculations, ensuring the scientific validity and reliability of the inversion calculations.

[0016] Based on the temperature infiltration data at each measuring point and the corresponding ambient temperature outside the wall, the temperature difference between the inside and outside of the wall at each measuring point is calculated. This temperature difference is the direct basis for heat flux density calculation, reflecting the magnitude of the driving force for heat transfer. The calculation process first acquires the temperature infiltration data at each measuring point, i.e., the measured values ​​from temperature sensors installed on the outer wall of the enclosure; simultaneously, the ambient temperature at the corresponding location is recorded, typically measured by a reference sensor located close to the enclosure but unaffected by its thermal radiation; then, the difference between the two is calculated to obtain the temperature difference between the inside and outside of the wall at each measuring point. For areas with insufficient measuring point density, an interpolation algorithm is used to generate a continuous temperature difference distribution field, ensuring coverage of the entire enclosure surface. The temperature difference data directly reflects the non-uniformity of heat transfer and hot spots, providing input conditions for heat flux density calculation.

[0017] Based on the temperature difference between the inside and outside of the wall, thermal conductivity, and wall thickness, the heat flux density at each measuring point corresponding to the location on the inner wall of the chamber is calculated using Fourier's law of heat conduction. Heat flux density is a physical quantity describing the rate of heat transfer per unit area and is a core parameter for retrieving the thermal field distribution. The calculation uses the classical Fourier law of heat conduction, with the following formula: ; in, The heat flux density is (W / m²). The thermal conductivity of the material is (W / (m·K)). The temperature difference (K) between the inside and outside of the wall. The value represents the wall thickness (m). The negative sign indicates that heat always flows from the high-temperature region to the low-temperature region.

[0018] For heterogeneous materials or complex structural walls, correction factors are used to adjust the calculation results, taking into account boundary effects and material nonlinear characteristics. The heat flux density calculation results form a heat flux distribution map of the inner wall of the box, which intuitively shows the spatial distribution characteristics of heat transfer and provides boundary conditions for the inversion of the spatial thermal field.

[0019] Based on the heat flux density at each location on the inner wall of the enclosure and the spatial distance from each location to each region inside the enclosure, a spatial thermal field distribution is generated using the heat source superposition method. The spatial thermal field distribution is a temperature distribution function in three-dimensional space, directly reflecting the thermal environment inside the enclosure. The inversion process is based on the principle of heat source superposition, treating the heat flux at each point on the inner wall of the enclosure as a heat source, calculating and superimposing the contributions of these heat sources to the temperature at each point in space. Specifically, a discretization method is used, dividing the internal space of the enclosure into a three-dimensional grid. For each grid point, the temperature contribution from all heat sources on the inner wall is accumulated. The temperature contribution calculation considers heat source intensity (heat flux density), distance attenuation, and spatial obstruction factors. The basic formula is: ; in, For spatial location The temperature at that location For reference base temperature, For the first The heat flux density of a heat source For the first The location of the heat source Let be the heat transfer influence function, representing the effect of a unit heat source on a distance of . The contribution of temperature.

[0020] For complex spatial layouts, the influence function Factors such as equipment obstruction, airflow, and radiative heat transfer were also considered, and coefficient corrections were used to reflect the heat transfer characteristics in the actual environment. The final generated spatial thermal field distribution is a three-dimensional temperature field function, which is visualized as temperature isosurfaces or slice maps, intuitively showing the temperature state of each region inside the enclosure and providing a thermal environment basis for subsequent analysis.

[0021] In this embodiment of the invention, the detailed implementation steps for calibrating and correcting the spatial thermal field distribution using infrared thermal imaging data include: The measured temperature distribution on the surface of visible equipment inside the enclosure is extracted from infrared thermal imaging data. This measured temperature distribution serves as the baseline data for calibrating the thermal field model, providing reliable temperature reference values. The extraction process first preprocesses the infrared thermal imaging data, including noise filtering, radiometric correction, and geometric calibration. Then, image segmentation techniques are used to identify the contours and boundaries of each device. Finally, the temperature matrix of the device surface is extracted, recording the temperature value corresponding to each pixel. For complex equipment, multi-angle imaging technology is employed to synthesize thermal image data from multiple perspectives to generate a complete surface temperature distribution. The measured temperature data possesses high spatial resolution and accuracy, precisely reflecting hot spots and temperature gradients on the device surface, providing a reliable reference standard for model calibration.

[0022] The calculated temperature values ​​at the corresponding equipment locations in the inverted spatial thermal field distribution are compared point-by-point with the measured temperature distribution to calculate the temperature deviation. The temperature deviation is a direct indicator of the model's accuracy, reflecting the difference between theoretical calculations and actual measurements. The comparison process first locates the precise positions of each device in the infrared thermal imaging in three-dimensional space; then, the calculated temperature values ​​at the corresponding locations are extracted from the spatial thermal field distribution; finally, the calculated temperature values ​​are compared point-by-point with the measured temperature values ​​to calculate the temperature deviation. Deviation calculation uses both absolute temperature difference and relative percentage methods to comprehensively evaluate the model's accuracy. For large equipment, the deviation analysis also considers spatial distribution characteristics, identifies systematic biases and random errors, provides a detailed difference analysis for model calibration, and clarifies the areas and parameters requiring key adjustments.

[0023] When the temperature deviation exceeds the preset allowable range, the boundary condition parameters of the inversion model are iteratively adjusted based on the measured temperature distribution. Boundary condition parameter adjustment is a crucial step in improving model accuracy, reducing the difference between theoretical calculations and measured values ​​through parameter optimization. The adjustment process employs an iterative optimization method. First, the objective function is defined as the weighted sum of squares of the temperature deviation. Then, the key parameters with the greatest impact on the deviation are identified, including material thermal conductivity, convective heat transfer coefficient, and heat source intensity. Finally, optimization methods such as gradient descent or genetic algorithms are used to iteratively adjust the parameter values ​​and minimize the objective function. To avoid overfitting, regularization constraints are introduced during the adjustment process to ensure the physical rationality of the parameter adjustments. For systematic deviations in different regions, a partitioned parameter adjustment strategy is adopted to specifically optimize local model parameters and improve the overall fitting effect.

[0024] The adjusted inversion model is recalculated to determine the spatial thermal field distribution until the temperature deviation converges to a preset tolerance range, resulting in the calibrated spatial thermal field distribution. Calibration correction is the final step in the inversion process, ensuring the accuracy and reliability of the final thermal field model. The correction process involves recalculating the thermal field based on the adjusted parameters, using the same heat source superposition principle but optimized boundary conditions and material parameters. The result is then compared again with the measured temperature to evaluate the new deviation. If the deviation still exceeds the tolerance range, parameter adjustments and recalculation continue, forming a closed-loop optimization process. When the deviation converges to a preset range (typically ±3℃ or ±5%), the model calibration is considered complete, and the final spatial thermal field distribution result is output. The calibrated thermal field distribution combines the complete coverage of the theoretical model with the accuracy of the measured data, providing a high-precision thermal environment foundation for subsequent thermal interaction analysis.

[0025] In this embodiment of the invention, the detailed implementation steps for calculating the thermal flux between each device and its adjacent devices based on the calibrated and corrected spatial thermal field distribution include: Local temperature values ​​at spatial locations of each device surface are extracted from the calibrated and corrected spatial thermal field distribution. These local temperature values ​​are fundamental data for thermal interaction analysis, reflecting the thermal environment state of the device surface. The extraction process first determines the precise location of the device surface within the spatial grid based on the device's three-dimensional geometric model; then, it interpolates and calculates the temperature values ​​at these locations from the calibrated spatial thermal field distribution; finally, it generates a temperature distribution matrix for the device surface. For devices with complex structures, high-density sampling points are used for fine extraction to ensure the capture of temperature gradients and hotspot regions. Local temperature data directly reflects the thermal environment conditions of the device, serving as the fundamental input for evaluating thermal interactions between devices and providing accurate temperature boundaries for subsequent temperature gradient calculations.

[0026] The process involves identifying the set of physically adjacent devices within the enclosure and calculating the spatial distance between each device and its neighbors. The set of physically adjacent devices defines the scope of thermal interaction analysis, and spatial distance is a key factor in heat transfer intensity. The process begins by identifying the spatial location and dimensions of each device based on the equipment layout diagram and 3D model of the prefabricated substation. Then, spatial proximity analysis is used to determine the set of physically adjacent devices for each device, typically employing a combination of distance thresholds and visibility assessment. Finally, the shortest and average spatial distances between devices are calculated as distance parameters for thermal interaction calculations. For devices with complex shapes, the spatial distance calculation considers the geometric characteristics of the actual heat dissipation surface, adopting the concept of equivalent heat dissipation distance and comprehensively considering area weighting and the main heat flow path to obtain a more accurate representation of the thermal interaction distance.

[0027] The temperature gradient between devices is calculated based on the difference in local temperature between each device and its adjacent devices, as well as the spatial distance between them. The temperature gradient is the driving force of heat transfer, directly determining the direction and intensity of heat flow. The calculation uses the definition of spatial temperature gradient, with the following formula: ; in, The temperature gradient is (K / m). and These represent the average temperature (K) of the surfaces of adjacent devices. The spatial distance between equipment (m).

[0028] For large equipment with uneven temperature distribution, a zoned calculation strategy is adopted. The equipment surface is divided into multiple regions, and the temperature gradient between each region and its adjacent equipment is calculated separately. Then, a weighted average is taken to obtain the overall gradient. The sign of the temperature gradient indicates the direction of heat flow, and the magnitude of the gradient reflects the potential intensity of heat interaction, providing a direct input for calculating heat interaction flux.

[0029] Based on the temperature gradient and the equivalent thermal conductivity of the air medium between devices, the heat exchange flux between each device and its adjacent devices is calculated. A positive heat exchange flux indicates that the device dissipates heat outward, while a negative value indicates that the device absorbs heat from adjacent devices. Heat exchange flux is a core indicator for quantifying heat exchange between devices, directly reflecting the intensity and direction of actual heat interaction. The calculation is based on the fundamental principles of heat transfer, comprehensively considering conduction, convection, and radiation. The formula is: ; in, For heat exchange flux (W). The equivalent thermal conductivity of the air medium in the equipment room (W / (m·K)). The effective area of ​​thermal interaction (m²). The temperature gradient is (K / m).

[0030] Equivalent thermal conductivity It is a comprehensive parameter that incorporates the combined effects of air heat conduction, convective heat transfer, and radiative heat transfer. The calculation formula is: ; in, The thermal conductivity of air, The convective heat transfer coefficient is... The radiative heat transfer coefficient is... For equipment spacing.

[0031] The calculated heat exchange flux forms a heat exchange matrix between devices, clearly showing the heat exchange network between the devices within the enclosure, providing a quantitative description of the heat exchange between devices for comprehensive heat load calculation. The positive and negative signs of the heat exchange flux intuitively reflect the direction of heat flow, helping to identify heat source and heat sink devices, which is of great significance for understanding the heat balance inside the enclosure.

[0032] In this embodiment of the invention, the detailed implementation steps for calculating the self-heating power of each device based on the load current data of each device include: Obtain the type identification and rated parameters of each piece of equipment. The type identification includes at least one of transformers, switches, busbars, and connectors. Type identification and rated parameters are fundamental information for self-heating calculations, determining the selection of the heating model and parameter settings. The acquisition process first extracts equipment information from the equipment list and technical specifications of the prefabricated substation; then, an equipment database is established, containing key parameters such as the type identification, rated current, rated voltage, rated power, and equipment specifications for each type of equipment; finally, parameters are corrected based on actual operating years and maintenance records, considering the impact of equipment aging. For transformers, key parameters recorded include rated capacity, no-load loss, and load loss; for switches, key parameters include rated current and contact resistance; for busbars and connectors, key parameters include material, cross-sectional area, and length. This fundamental information provides the necessary conditions for subsequently selecting a suitable heating model and calculating accurate heating power.

[0033] Real-time load current values ​​of each device are collected, and the load factor relative to the rated current is calculated. Real-time load current and load factor are direct indicators of equipment operating status, determining the actual heat generation level of the equipment. The data acquisition process involves continuously recording the operating current of each device using current transformers and monitoring systems installed in key locations; then, signal processing is performed, including filtering, calibration, and digitization; finally, the load factor, i.e., the ratio of real-time current to rated current, is calculated. For three-phase equipment, the three-phase current is collected separately, taking into account the impact of three-phase imbalance. The load factor is a standardized representation of the equipment's operating status, directly reflecting the equipment's load level and is a key input parameter for calculating self-heating power.

[0034] Select the corresponding heating model based on the type identifier. For transformers, calculate the copper and iron loss components separately; for switches, calculate the contact resistance loss component; and for conductors, calculate the resistance loss and transition resistance loss components. The heating model is the theoretical basis for calculating equipment self-heating, accurately estimating various losses based on equipment characteristics. For transformers, copper loss is calculated based on the square of the load current and winding resistance, while iron loss is calculated based on the core material properties and magnetic flux density. For switches, contact resistance loss is calculated based on the contact surface resistance and the square of the current flowing through it. For conductors, resistance loss is based on the conductor's body resistance, and transition resistance loss is based on the contact resistance at the connection point. These heating models are all based on electrical engineering and thermodynamics principles, accurately describing the heating mechanisms of different types of equipment through physical formulas, providing a theoretical basis for calculating heating power.

[0035] The current operating temperature of each device is extracted from the spatial thermal field distribution. A resistance temperature rise correction factor is calculated based on the temperature coefficient of the resistivity of the conductive material and applied to each loss component. Resistance temperature rise correction is a key step in improving the accuracy of heat generation calculations, taking into account the influence of temperature on resistance. The correction process first extracts the average operating temperature of the devices from the spatial thermal field distribution; then, it calculates the change in resistance with temperature based on the temperature coefficient of the material's resistance; finally, it applies the correction factor to adjust each loss component. For copper conductors, the formula for calculating the resistance temperature rise correction factor is: ; in, This is the resistance temperature rise correction factor. This is the temperature coefficient of resistance of copper (approximately 0.00393 / ℃). The current operating temperature. This is a reference temperature (usually 20℃).

[0036] After applying the correction factor, each loss component more accurately reflects the heating status under actual temperature conditions, thus improving the accuracy of self-heating power calculation.

[0037] The historical load current sequence of each device is obtained, and the load fluctuation coefficient within a preset time window is calculated. When the load fluctuation coefficient exceeds a preset threshold, an additional fluctuation loss component is added. Load fluctuation loss is an important supplement to conventional heating models, reflecting the additional thermal stress caused by dynamic load changes. The calculation process first obtains the load current sequence within a preset time window (usually 15-30 minutes); then, the load fluctuation coefficient is calculated, usually defined as the ratio of the current standard deviation to the average value; when the fluctuation coefficient exceeds the threshold (usually 0.2-0.3), the device is considered to be in a state of significant load fluctuation, and additional losses need to be considered. The fluctuation loss calculation usually uses empirical formulas, which are related to the fluctuation frequency and amplitude, reflecting the additional thermal stress and losses caused by frequent load changes to the device. This step is particularly important because traditional heating models are usually based on steady-state assumptions and cannot accurately reflect the actual heating situation under dynamic load conditions.

[0038] The basic self-heating power is obtained by summing all loss components of each device. Then, an aging and deterioration correction factor is introduced based on the device's operating years to make an upward adjustment, resulting in the final self-heating power for each device. The final self-heating power calculation is a comprehensive process that considers the device's basic losses and various correction factors. The summation process adds the previously calculated loss components to obtain the basic self-heating power; then, an aging and deterioration correction factor is introduced to make an upward adjustment based on the device's operating years and condition. The aging correction factor is usually determined based on the device's lifespan curve model, reflecting the performance degradation and increased losses caused by prolonged use. For example, for equipment that has been used for more than half of its rated lifespan, the correction factor may be between 1.1 and 1.3, with the specific value determined based on the equipment type and maintenance status. The final calculated self-heating power is a comprehensive indicator that comprehensively considers equipment type, operating status, temperature influence, load fluctuations, and aging degree, accurately reflecting the equipment's heating status under actual operating conditions and providing reliable heat source data for subsequent heat load analysis.

[0039] In this embodiment of the invention, the detailed implementation steps of superimposing the heat exchange flux and the self-heating power to obtain the comprehensive heat load value of each device, and calculating the thermal stress accumulation rate of each device based on the time-series change of the comprehensive heat load value, include: The comprehensive heat load value of each device at the current moment is obtained by subtracting the sum of its heat exchange flux with all adjacent devices from its self-heating power. The comprehensive heat load value is a key indicator of the actual heat accumulation of the equipment, reflecting the balance between heat generation and dissipation. The calculation process first summarizes the self-heating power of the equipment as the heat generation item; then it summarizes the heat exchange flux with all adjacent devices, noting the sign of the heat exchange flux—positive values ​​indicate heat dissipation, and negative values ​​indicate heat absorption; finally, algebraic operations are performed to obtain the comprehensive heat load value. The calculation formula is: ; in, The comprehensive heat load value, Self-heating power, It is the sum of thermal flux with all adjacent devices.

[0040] A positive value for the comprehensive heat load indicates that the equipment is in a state of heat accumulation, while a negative value indicates that it is in a state of heat dissipation. This indicator directly reflects the real-time thermal balance of the equipment and provides basic data for thermal stress assessment.

[0041] The comprehensive heat load value sequence of each device is recorded within a continuous monitoring period. This comprehensive heat load value sequence forms the time-series data basis for thermal stress analysis, reflecting the dynamic changes in the thermal state of the equipment. The recording process employs a timed sampling strategy, typically with sampling intervals of 5-15 minutes, continuously recording the comprehensive heat load values ​​of each device throughout the complete monitoring period (usually 24 hours or 7 days). Data preprocessing is then performed, including outlier detection, missing value completion, and smoothing filtering. Finally, a standardized time-series dataset is formed, containing timestamps and corresponding comprehensive heat load values. For scenarios with complex operating modes, load change events and external environmental changes also need to be recorded to help explain the reasons for load sequence fluctuations. The complete comprehensive heat load value sequence provides a comprehensive temporal perspective of the equipment's thermal state, offering reliable time-series data for thermal stress accumulation analysis.

[0042] The comprehensive heat load value series is processed by time-series integration to calculate the total cumulative heat load of each device during the monitoring period. The total cumulative heat load is a key indicator for assessing long-term thermal effects, quantifying the total heat accumulation within the monitoring period. The integration process employs numerical integration methods, typically using the trapezoidal rule or Simpson's rule, to calculate the area under the time series curve. The integration results are then normalized to eliminate the influence of the monitoring period length, finally yielding a standardized total cumulative heat load. For negative values ​​(where heat dissipation exceeds heat generation), only the positive values ​​can be integrated for a more accurate reflection of the heat accumulation effect. The total cumulative heat load directly reflects the heat accumulation level of the equipment throughout the monitoring period and is an important intermediate result for thermal stress assessment.

[0043] A baseline value for thermal stress is established, and the portion of the total accumulated heat load exceeding this baseline value is divided by the monitoring duration to obtain the thermal stress accumulation rate for each piece of equipment. The thermal stress accumulation rate is a core indicator for quantifying the rate of thermal damage to equipment, reflecting the long-term impact of heat loads exceeding safety thresholds on the equipment. The calculation process first sets an appropriate baseline value for thermal stress based on the equipment type and material properties, typically based on the heat load level corresponding to the equipment's long-term allowable operating temperature. Then, the total accumulated heat load is subtracted from the product of the baseline value and the monitoring duration to obtain the excess heat load accumulation. Finally, this is divided by the monitoring duration to obtain the thermal stress accumulation rate per unit time. This method ensures that only heat loads exceeding safety thresholds are included in the thermal stress calculation, more accurately reflecting the actual thermal damage process. The unit of thermal stress accumulation rate is power (W), which intuitively represents the rate of thermal damage and provides a key input parameter for insulation life assessment.

[0044] In this embodiment of the invention, the detailed implementation steps for setting a thermal stress starting point value and dividing the portion of the total accumulated heat load exceeding the thermal stress starting point value by the monitoring duration to obtain the thermal stress accumulation rate of each device include: Based on the long-term allowable operating temperature of the insulation materials of each piece of equipment, and combined with the normal heat generation power under the rated load, the starting benchmark value for thermal stress calculation is calculated for each piece of equipment. The starting benchmark value for thermal stress calculation is a safe threshold for thermal stress calculation, reflecting the level of heat load that the equipment can withstand over a long period. The calculation process first obtains the long-term allowable operating temperature of the equipment's insulation materials, usually from insulation material specifications and equipment design standards; then, using the heat balance equation, the corresponding heat load level at the allowable temperature is calculated, considering the heat balance state under normal heat dissipation conditions; finally, a specific benchmark value is determined as the starting point for thermal stress calculation. The physical meaning of the benchmark value is the balanced heat load of the equipment under normal operating conditions; only heat accumulation exceeding this level will lead to accelerated insulation aging and performance degradation. The rationality of the benchmark value setting directly affects the accuracy of thermal stress assessment, and it is usually verified and adjusted through comparison with historical operating data and equipment condition.

[0045] The accumulated excess heat load is obtained by subtracting the product of the thermal stress baseline value and the monitoring duration from the total accumulated heat load. The accumulated excess heat load represents the effective heat actually causing thermal damage, excluding the basic heat load that the equipment can safely withstand. The calculation formula is: ; in, This is the cumulative amount of excess heat load. To accumulate the total heat load, This serves as the baseline value for calculating thermal stress. For monitoring duration.

[0046] This calculation method is based on the threshold effect theory of thermal damage, which posits that only thermal loads exceeding the material's heat resistance will lead to substantial damage. The cumulative amount of excess thermal load directly reflects the actual amount of thermal damage the equipment experiences during the monitoring period, serving as the direct basis for calculating the thermal stress rate.

[0047] When the accumulated excess heat load is negative, the thermal stress accumulation rate of the corresponding equipment is set to zero. Negative value processing is a logical protection mechanism for thermal stress calculation, ensuring the physical rationality of the calculation results. When the accumulated excess heat load is negative, it indicates that the actual heat load of the equipment is below the safety threshold, it is in normal operating condition, and no thermal damage will occur; therefore, the thermal stress accumulation rate should be zero. This processing rule is based on the physical fact that thermal damage is irreversible and has a threshold characteristic, ensuring that the thermal stress assessment results conform to the actual damage mechanism. In engineering practice, different negative value processing strategies can be set according to the equipment type and importance. For example, for critical equipment, a small amount of basic thermal stress can be retained to reflect the cumulative effect of long-term operation.

[0048] When the accumulated excess heat load is positive, the accumulated excess heat load is divided by the monitoring duration to obtain the thermal stress accumulation rate of the equipment. The calculation of the thermal stress accumulation rate is a standardized step that converts the accumulated amount into a rate, facilitating comparison of results from different monitoring periods and long-term trend analysis. The calculation formula is: ; in, The rate of thermal stress accumulation. This is the cumulative amount of excess heat load. For monitoring duration.

[0049] The thermal stress accumulation rate is measured in power (W), and its physical meaning is the effective heat that causes damage to equipment per unit time, directly reflecting the rate at which thermal damage occurs. This indicator can be directly used to assess the thermal health status of equipment, predict the aging process of insulation materials, and provide a quantitative basis for subsequent lifespan analysis. In practical applications, the changing trend of the thermal stress accumulation rate can also be calculated to assess the rate of equipment deterioration, providing more reference information for preventive maintenance decisions.

[0050] In this embodiment of the invention, the detailed implementation steps for calculating the remaining insulation life consumption rate of each device based on the thermal stress accumulation rate and the thermal degradation sensitivity coefficient of the insulation material of each device include: The heat resistance rating and corresponding thermal degradation sensitivity coefficient of the insulation materials for each piece of equipment are obtained. The thermal degradation sensitivity coefficient characterizes the degree of insulation life loss due to the cumulative amount of thermal stress per unit. It is a key parameter connecting thermal stress and lifespan, reflecting the sensitivity of different materials to thermal damage. The process begins by determining the type and heat resistance rating of the insulation materials for each piece of equipment, typically extracted from equipment specifications and material standards. Then, the corresponding thermal degradation sensitivity coefficient is determined through querying or testing. This coefficient is usually obtained through accelerated aging tests and represents the relative rate of material lifespan loss under unit thermal stress. Insulation materials with different heat resistance ratings have different sensitivity coefficients; generally, the higher the heat resistance rating, the smaller the sensitivity coefficient, indicating stronger resistance to thermal damage. The accuracy of the sensitivity coefficient directly affects the reliability of lifespan assessment. Therefore, in practical applications, it is often corrected and optimized by combining historical operating data and equipment condition monitoring results to ensure that the coefficient value matches the actual aging rate.

[0051] The thermally induced insulation loss component is obtained by multiplying the cumulative rate of thermal stress for each piece of equipment by its corresponding thermal degradation sensitivity coefficient. This component represents the lifespan attrition rate caused by thermal factors, reflecting the direct impact of thermal stress on insulation aging. A linear model is used in the calculation, assuming a proportional relationship between thermal stress and lifespan attrition within a certain range. The physical unit of the thermally induced loss component is typically "standard lifespan percentage / time," which intuitively represents the percentage of standard lifespan lost per unit time due to thermal factors. This component is a major part of the remaining insulation lifespan attrition rate, directly reflecting the degree of impact of thermal aging on equipment lifespan and providing fundamental data for overall lifespan assessment.

[0052] Obtain the current load rate of each device. When the load rate exceeds a preset percentage of the rated load, calculate the overload additional loss component. Overload additional loss is a supplementary factor considering electrical factors, reflecting the extra damage to insulation caused by high-load operation. The calculation process first compares the current load rate with a preset threshold (usually 90% or 95% of the rated load) to determine if the equipment is overloaded. Then, it calculates the additional loss coefficient based on the degree of overload, typically using a nonlinear model to reflect the characteristic that overload damage increases sharply with the degree of overload. Finally, it derives the overload additional loss component as a supplementary electrical factor to the insulation life consumption. Overload additional loss considers comprehensive factors such as increased electric field strength, increased mechanical stress, and soaring hotspot temperatures, and is an important extension of conventional thermal aging models. Especially for equipment that frequently operates under overload, the impact of this component cannot be ignored.

[0053] Specifically, the calculation process for the overload additional loss component is as follows: The overload additional loss coefficient is calculated using a nonlinear model. This reflects the characteristic that overload damage increases sharply with the degree of overload, and is usually expressed in the following form: ; in, The current load rate of the equipment. The preset load threshold (usually 90% or 95% of the rated load). This is the overload sensitivity coefficient, which is related to the type of equipment and material properties; The nonlinear exponent, typically greater than 1, characterizes the nonlinear growth characteristics of overload damage; Operating time influence: Considering the effect of overload duration t, a time-weighted coefficient is calculated. ; Final calculation: The overload additional loss component is calculated as follows: ; in, Added loss component for overload; This is the baseline loss rate, which is usually related to the basic parameters of the equipment. This is the additional loss factor for overload. The thermally induced insulation loss component is added to the overload-induced additional loss component to obtain the insulation remaining life consumption rate of each piece of equipment. The insulation remaining life consumption rate is the final indicator for comprehensively assessing the aging rate of equipment, integrating the combined effects of thermal and electrical factors. The calculation uses a simple linear superposition model, assuming that the damage effects caused by different factors can be independently accumulated. The unit of consumption rate is usually expressed as "standard life percentage / year," which intuitively reflects the percentage of standard life consumed by the equipment insulation each year under current operating conditions. This indicator is a core quantitative indicator for equipment health assessment, directly related to early warning decisions and maintenance planning, and provides a basic criterion for subsequent identification of target equipment for early warning. In practical applications, the consumption rate can be compared with historical trends and similar equipment to further assess the relative deterioration and abnormal level of equipment condition, improving the targeting and accuracy of early warnings.

[0054] In this embodiment of the invention, the detailed implementation steps for identifying and issuing early warning target equipment based on the remaining insulation life consumption rate and the series connection position of each device in the high and low voltage side electrical circuits of the prefabricated substation include: A topology diagram of the main electrical pathway from the high-voltage incoming line side to the low-voltage outgoing line side of a prefabricated substation is constructed. This main electrical pathway topology diagram is a network representation of the system structure, intuitively reflecting the connection relationships of equipment and the power supply path. The construction process first obtains the electrical schematic diagram and equipment layout diagram of the prefabricated substation; then, it identifies the main power supply paths, including key equipment such as the high-voltage incoming switch, high-voltage busbar, transformer, low-voltage busbar, and outgoing switch; finally, it establishes a standardized topology diagram structure, clearly showing the connection relationships of equipment and the power flow. The topology diagram uses a directed graph data structure, where nodes represent equipment, directed edges represent power flow, and edge weights can represent line capacity or importance. For complex structures, the topology diagram also needs to label parallel paths and backup equipment to comprehensively reflect the system's redundancy design and switching capabilities. An accurate topology structure is the foundation for assessing equipment importance and the impact of faults, providing a system perspective for early warning target identification.

[0055] In the electrical main path topology diagram, each device node is labeled, and the upstream and downstream devices of each device are marked according to the power supply direction. Device relationship labeling is a fundamental step in topology analysis, clarifying the position of devices in the power supply chain and their interdependencies. The labeling process first determines the power flow direction, typically from the high-voltage side to the low-voltage side; then, for each device node, its directly connected upstream devices (power supply side) and downstream devices (load side) are identified; finally, a complete connection relationship database is established, recording the upstream and downstream device sets for each device. For dual-power or ring network structures, the variability of power flow direction must also be considered, labeling multiple possible power supply paths. Clear upstream and downstream relationships help analyze the cascading effects and impact range of device failures, provide a network perspective for identifying critical devices, and ensure that the selection of early warning targets takes into account system structural factors.

[0056] Equipment whose insulation remaining life consumption rate exceeds a preset consumption threshold is identified as abnormal consumption equipment. Screening abnormal consumption equipment is the first step in identifying high-risk equipment, and preliminary screening is based on quantitative indicators. The screening process first determines an appropriate preset consumption threshold, typically based on industry standards and historical experience, such as 5% or 10% of the standard life per year. Then, the insulation remaining life consumption rate of each piece of equipment is compared with the threshold to identify equipment exceeding the threshold. Finally, a list of abnormal consumption equipment is compiled, including key information such as equipment identification, location, consumption rate, and the extent of exceeding the threshold. Threshold settings can be differentiated according to equipment type and importance, with stricter standards applied to critical equipment. Abnormal consumption equipment is a potential high-risk target, but considering only the equipment's own condition, further screening based on system structure is necessary to determine the true warning targets.

[0057] The process involves determining whether abnormally consuming equipment has a parallel backup branch in the main electrical path topology diagram. Equipment without a parallel backup branch is marked as a warning target. Backup path determination is a crucial step in ultimately identifying warning targets, considering the system impact and substitutability of equipment failures. The process first searches the topology diagram for other equipment paths with the same upstream and downstream nodes as the abnormally consuming equipment; then, it assesses the operating status and switching capabilities of these paths to determine if they are valid backup paths; finally, it filters out abnormally consuming equipment without valid backup paths and marks them as the final warning target equipment. This step ensures the targeted nature and effectiveness of the warning, focusing on critical equipment with abnormal states and lacking redundant protection. Failure of these devices will directly lead to system shutdown, giving them higher warning value and greater necessity for intervention. The final list of warning target equipment forms the basis for subsequent warning signal generation, directly determining the scope and focus of the warning.

[0058] In this embodiment of the invention, the detailed implementation steps for determining whether an abnormally consuming device has a parallel backup branch in the electrical main path topology diagram include: In the electrical main path topology diagram, other equipment branches sharing the same upstream power supply node and downstream power receiving node as the abnormally consuming equipment are identified. Parallel branch search is a fundamental step in identifying redundant paths, discovering potential backup paths through topology analysis. The search process first determines the exact location and connection relationship of the abnormally consuming equipment, identifying its upstream power supply node and downstream power receiving node; then, it searches the topology diagram for other equipment or combinations of equipment with the same connection endpoints, which constitute potential parallel branches; finally, it generates a list of alternative paths, recording the constituent equipment and connection methods of each path. For complex network structures, breadth-first search or depth-first search algorithms are used to improve search efficiency, ensuring that all possible alternative paths are discovered. The search results directly reflect the system's redundancy design level, providing fundamental data for backup capability assessment.

[0059] If other equipment branches exist, the current operating status and switching capability of each device in those branches are detected. Operating status detection is a crucial step in assessing the effectiveness of the backup path, ensuring that backup equipment can take over when needed. The detection process first acquires real-time status information of each device in the backup branch, including key indicators such as operating status (running / standby / disabled), health status, and load level; then, it assesses switching capability, including switching time, automation level, and switching reliability; finally, it generates a status assessment report for the backup branch, comprehensively reflecting its current availability and responsiveness. For manually switched equipment, operator arrival time and operational complexity must also be considered; for automatically switched equipment, the focus is on assessing the reliability and response speed of the control system. The status detection results directly affect the judgment of the backup path's effectiveness; only backup equipment in good condition and capable of rapid deployment can provide true redundancy protection.

[0060] When other equipment branches are in hot standby mode and can complete load transfer within a preset switching time, the abnormally consuming equipment is determined to have a parallel backup branch. Effective backup determination is the final decision in backup path assessment, determining whether the abnormal equipment has practically usable backup protection. The determination criteria mainly consider two key factors: first, the standby status—equipment in hot standby (powered but not loaded) state can quickly take over the load, providing higher reliability; second, the switching time—load transfer must be completed within a preset time limit (usually in the range of seconds or minutes, depending on the importance of the load) to ensure power continuity. Only backup paths that simultaneously meet both conditions are considered effective, and the corresponding abnormally consuming equipment is determined to have a parallel backup branch. This determination directly affects the screening of early warning targets; equipment with effective backup paths usually has lower priority unless its state is extremely abnormal or the backup path also has potential risks.

[0061] When no other equipment branch meets the conditions, the abnormal power consumption device is determined to lack a parallel backup branch and is marked as a warning target device. Final warning target confirmation is the last step in the warning object screening process, clarifying the devices requiring key attention and intervention. When an abnormal power consumption device lacks an effective parallel backup path, it means that a failure of this device will directly lead to a power outage, lacking a redundant protection mechanism; therefore, it is marked as a warning target device. Such devices have dual risk characteristics: firstly, their own condition is abnormal, with excessively rapid insulation aging; secondly, the system structure is fragile and lacks effective backup. The confirmation of warning target devices comprehensively considers both device status and system structure dimensions, ensuring the relevance and value of the warning and providing a clear set of objects for subsequent generation of tiered warning signals.

[0062] In this embodiment of the invention, the detailed implementation steps for generating graded early warning signals based on the estimated remaining insulation life of the target equipment and the outage load capacity of the prefabricated substation caused by its failure include: The initial insulation life baseline value and historical cumulative consumption of each early warning target device are obtained to calculate the remaining insulation life margin. The remaining insulation life margin is a time-dimensional risk indicator, reflecting the expected remaining time before equipment failure. The calculation process first obtains the initial insulation life baseline value of the equipment, typically extracted from industry standards or manufacturer specifications based on equipment type and insulation class; then, the historical cumulative consumption is obtained, and the percentage of cumulatively consumed insulation life is calculated using historical operating data and status records; finally, the remaining life margin is calculated, which is the initial life margin minus the consumed life. For equipment with a long service life, the nonlinear aging effect must also be considered, and the historical consumption is adjusted using an acceleration factor to more accurately reflect the actual aging state. The remaining insulation life margin is usually expressed as a percentage of standard life, intuitively reflecting the equipment's health reserve and providing basic data for failure time prediction.

[0063] Dividing the remaining insulation life margin by the remaining insulation life consumption rate yields the estimated remaining insulation life. Estimating remaining insulation life is a crucial step in time-of-failure forecasting, transforming the remaining life and consumption rate into specific time projections. The calculation employs a simple linear extrapolation model, assuming the current consumption rate remains constant over a future period. The formula for calculating the estimated life is: ; in, This is an estimate of the remaining life of the insulation. This represents the remaining insulation life margin (percentage of standard life). This represents the remaining insulation life consumption rate (standard life percentage / year).

[0064] The estimated value is in years, directly reflecting how long the equipment is expected to operate safely under the current operating conditions. This indicator is a key basis for assessing the urgency of risks; the smaller the estimated value, the more urgent the risk, requiring a higher level of early warning and faster intervention. In practical applications, historical trend analysis and probability models can be combined to provide more refined forecast ranges and credibility assessments, improving the reliability and accuracy of the forecast.

[0065] The outage load capacity is obtained by traversing all downstream load nodes of each early warning target device in the electrical main path topology diagram and summing the rated capacities of the downstream load nodes. The outage load capacity is a quantitative indicator of the impact range, reflecting the degree of impact of equipment failure on the system. The calculation process first determines the precise location of the early warning target device in the electrical topology diagram; then, using a graph traversal algorithm (usually depth-first search), all downstream load nodes of the device are identified; finally, the rated capacities of these nodes are summed to obtain the total potential outage load. For load nodes with multiple power supply paths, the availability and switching capability of backup power must be considered, and only the portion that cannot be powered by backup paths is calculated. The outage load capacity is usually expressed in kilowatts (kW) or kilovolt-amperes (kVA), intuitively reflecting the scale of the fault's impact and is a key indicator for assessing the severity of risk. The larger the capacity, the wider the impact range, the higher the risk level, and the higher the priority for attention and handling.

[0066] The urgency level is determined based on the estimated remaining insulation life, and the impact level is determined based on the proportion of out-of-service load capacity to the total capacity of the prefabricated substation. Risk level classification is the fundamental step in early warning grading, transforming continuous risk indicators into discrete level classifications. Urgency levels are typically based on time thresholds, for example: less than 1 year is extremely urgent, 1-3 years is high urgent, 3-5 years is medium urgent, and more than 5 years is low urgent. Impact levels are based on capacity proportions, for example: above 50% is extremely high impact, 30%-50% is high impact, 10%-30% is medium impact, and below 10% is low impact. The classification standards can be adjusted according to actual application scenarios and management requirements to ensure the practicality and applicability of early warning grading. Clear level definitions help standardize the risk assessment process, provide a unified way of expressing risk, and facilitate comparison and management between different systems and scenarios.

[0067] The final warning level is determined by combining the urgency and impact levels, generating a tiered warning signal. The final warning level is the result of a comprehensive risk assessment, integrating both time urgency and impact severity. The determination process typically employs a risk matrix method, mapping the urgency and impact levels to a unified warning level framework, such as four levels: Red (requiring immediate action), Orange (requiring priority action), Yellow (requiring planned action), and Green (requiring monitoring). The matrix mapping rule usually follows the "highest priority" principle, taking the higher level from the two dimensions as the final level, or calculating a comprehensive score through a weighted average. The generation of the warning signal includes three parts: level determination, detailed description, and handling recommendations, forming a complete warning information package. This package is transmitted to the power distribution operation and maintenance system through a standard interface to support subsequent decision-making and intervention. The establishment of a tiered warning system makes risk management more scientific and refined, enabling the optimization of resource allocation and intervention strategies based on risk characteristics, and improving the overall practicality and effectiveness of the warning system.

[0068] This invention achieves comprehensive monitoring and early warning of the operating status of prefabricated substations by collecting multimodal data, spatial thermal field inversion, thermal interaction analysis, thermal stress calculation, insulation life assessment, and graded early warning output. The multimodal data fusion method of this invention can accurately assess the thermal state and lifespan of equipment, effectively identify high-risk equipment, and provide a systematic solution for the safe operation and preventative maintenance of prefabricated substations.

[0069] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0070] It should be noted that all formulas in this manual are calculated by removing dimensions and taking their numerical values. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0071] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for early warning of the operating status of a prefabricated substation based on multimodal data, characterized in that, include: Infrared thermal imaging data and load current data of each device inside the sealed box of the prefabricated substation were collected, as well as temperature penetration data of multiple measuring points on the outer wall of the box. Based on the temperature penetration data and the thermal conductivity characteristics of the box wall material, the spatial thermal field distribution of each area inside the box is calculated by inversion, and the spatial thermal field distribution is calibrated and corrected by combining the infrared thermal imaging data. Based on the calibrated and corrected spatial thermal field distribution, the heat exchange flux between each device and adjacent devices is calculated, and the self-heating power of each device is calculated based on the load current data of each device. The heat exchange flux is superimposed with the self-heating power to obtain the comprehensive heat load value of each device, and the thermal stress accumulation rate of each device is calculated based on the time-series change of the comprehensive heat load value. Based on the thermal stress accumulation rate and the thermal degradation sensitivity coefficient of the insulation material of each device, the insulation remaining life consumption rate of each device is calculated. Based on the insulation remaining life consumption rate and the series connection position of each device in the high and low voltage side electrical circuit of the prefabricated substation, the warning target device whose insulation remaining life consumption rate exceeds the preset consumption threshold and is located on the main power supply path is identified. Based on the estimated remaining insulation life of the target equipment and the outage load capacity of the prefabricated substation caused by its failure, a graded early warning signal is generated and output to the power distribution operation and maintenance system.

2. The method according to claim 1, characterized in that, The step of calculating the spatial thermal field distribution in each region inside the enclosure based on the temperature permeation data and the thermal conductivity characteristics of the enclosure wall material includes: Obtain the thermal conductivity and wall thickness parameters of each wall surface of the enclosure; Based on the temperature penetration data at each measuring point and the corresponding ambient temperature outside the wall, calculate the temperature difference between the inside and outside of the wall at each measuring point; Based on the temperature difference between the inside and outside of the wall, the thermal conductivity and the wall thickness parameters, the heat flux density at each measuring point corresponding to the inner wall position of the box is calculated using Fourier's law of heat conduction. Based on the heat flux density at each position on the inner wall of each box and the spatial distance from each position to each region inside the box, the spatial thermal field distribution is generated by inversion using the heat source superposition method.

3. The method according to claim 1, characterized in that, The calculation of heat exchange flux between each device and its adjacent devices based on the calibrated and corrected spatial thermal field distribution includes: Extract the local temperature value of each device surface at its spatial location from the calibrated and corrected spatial thermal field distribution; Determine the set of physically adjacent devices within the enclosure for each device, and calculate the spatial distance between each device and its adjacent devices; The temperature gradient between the devices is calculated based on the difference in local temperature values ​​between each device and its adjacent devices, as well as the spatial distance between them. Based on the temperature gradient and the equivalent thermal conductivity of the air medium between the devices, the heat exchange flux between each device and its adjacent devices is calculated.

4. The method according to claim 1, characterized in that, The step of superimposing the heat exchange flux with the self-heating power to obtain the comprehensive heat load value of each device, and calculating the thermal stress accumulation rate of each device based on the time-series change of the comprehensive heat load value, includes: The sum of the heat exchange flux between each device and all its adjacent devices is subtracted from the self-heating power of each device to obtain the comprehensive heat load value of each device at the current moment. Record the sequence of comprehensive heat load values ​​for each device during the continuous monitoring period; The comprehensive heat load value sequence is subjected to time-series integration processing to calculate the total cumulative heat load of each device during the monitoring period; A thermal stress starting point is set, and the portion of the total accumulated heat load that exceeds the thermal stress starting point is divided by the monitoring time to obtain the thermal stress accumulation rate of each device.

5. The method according to claim 1, characterized in that, The calculation of the remaining insulation life consumption rate of each device based on the thermal stress accumulation rate and the thermal degradation sensitivity coefficient of the insulation material of each device includes: Obtain the heat resistance rating and corresponding thermal degradation sensitivity coefficient of the insulation material of each device. The thermal degradation sensitivity coefficient characterizes the degree of loss of insulation life due to the unit thermal stress accumulation. The thermal stress accumulation rate of each device is calculated as a product of its corresponding thermal degradation sensitivity coefficient to obtain the thermally induced insulation loss component. Obtain the current load rate of each device, and when the load rate exceeds a preset proportion of the rated load, calculate the overload additional loss component; The thermally induced insulation loss component is added to the overload additional loss component to obtain the insulation remaining life consumption rate of each device.

6. The method according to claim 1, characterized in that, The method of identifying early warning target equipment whose insulation remaining life consumption rate exceeds a preset consumption threshold and is located on the main power supply path based on the insulation remaining life consumption rate and the series connection position of each device in the high and low voltage side electrical circuits of the prefabricated substation includes: Construct the electrical main path topology diagram of the prefabricated substation from the high-voltage incoming line side through the transformer to the low-voltage outgoing line side; Mark each device node in the electrical main path topology diagram, and mark the upstream and downstream devices of each device according to the power supply direction; Devices whose insulation remaining life consumption rate exceeds the preset consumption threshold are identified as abnormal consumption devices. Determine whether the abnormally consuming device has a parallel backup branch in the electrical main path topology diagram, and mark the abnormally consuming device that does not have a parallel backup branch as the early warning target device.

7. The method according to claim 1, characterized in that, The step of generating tiered early warning signals based on the estimated remaining insulation life of the target equipment and the resulting outage load capacity of the prefabricated substation after its failure includes: Obtain the initial insulation life baseline value and historical cumulative consumption of each of the aforementioned early warning target devices, and calculate the remaining insulation life margin. Divide the remaining insulation lifetime margin by the remaining insulation lifetime consumption rate to obtain the estimated remaining insulation lifetime. The rated capacity of each downstream load node of the warning target equipment is obtained by traversing all downstream load nodes in the electrical main path topology diagram and summing the rated capacity of the downstream load nodes. The urgency level is determined based on the estimated remaining insulation life, and the impact level is determined based on the proportion of the out-of-service load capacity to the total capacity of the prefabricated substation. The final warning level is determined by combining the urgency level and the impact level, and the graded warning signal is generated.

8. The method according to claim 2, characterized in that, The calibration and correction of the spatial thermal field distribution based on the infrared thermal imaging data includes: The measured temperature distribution on the surface of the visible device inside the box is extracted from the infrared thermal imaging data. The calculated temperature value corresponding to the equipment location in the inverted spatial thermal field distribution is compared point by point with the measured temperature distribution to calculate the temperature deviation value. When the temperature deviation exceeds the preset allowable range, the boundary condition parameters of the inversion model are iteratively adjusted based on the measured temperature distribution. The adjusted inversion model is used to recalculate the spatial thermal field distribution until the temperature deviation value converges to the preset deviation tolerance range, thus obtaining the calibrated and corrected spatial thermal field distribution.

9. The method according to claim 4, characterized in that, The process of setting a thermal stress starting point benchmark value, dividing the portion of the total accumulated heat load exceeding the thermal stress starting point benchmark value by the monitoring duration, yields the thermal stress accumulation rate for each device, including: Based on the long-term allowable operating temperature of the insulation material of each piece of equipment, and combined with the normal heat generation power under the rated load of the equipment, the starting base value of thermal stress for each piece of equipment is calculated. Subtracting the product of the thermal stress starting point value and the monitoring duration from the total accumulated heat load, we obtain the accumulated excess heat load. When the accumulated excess heat load is negative, the thermal stress accumulation rate of the corresponding equipment is set to zero; When the accumulated excess heat load is positive, the accumulated excess heat load is divided by the monitoring duration to obtain the thermal stress accumulation rate of the equipment.

10. The method according to claim 6, characterized in that, The determination of whether the abnormally consuming equipment has a parallel backup branch in the electrical main path topology diagram includes: In the electrical main path topology diagram, find other equipment branches that have the same upstream power supply node and the same downstream power receiving node as the abnormal power consumption device; If other equipment branches exist, detect the current operating status and switching capability of each device in the other equipment branches; When the other equipment branch is in hot standby mode and the load transfer can be completed within a preset switching time, it is determined that the abnormally consuming equipment has a parallel standby branch. When no other equipment branch meets the conditions, it is determined that the abnormally consuming equipment does not have a parallel backup branch, and it is marked as the warning target equipment.

Images