A power distribution box anti-overheating early warning method integrating a wireless module

Abstract

This invention discloses an overheating warning method for distribution boxes with integrated wireless modules, relating to the field of overheating warning technology for distribution boxes. The method includes: acquiring temperature signals and performing dynamic filtering and drift compensation processing; dividing the processed signals into regions and obtaining complete temperature field data through structurally constrained spatial interpolation to identify local hotspots; constructing a thermodynamic model for local hotspot regions and determining an adaptive temperature threshold by combining temperature safety margin and temperature rise trend correction; comparing real-time temperature data with the adaptive temperature threshold, triggering an overheating warning using a dual judgment mechanism of continuous over-threshold cumulative counting and duration, and conducting risk assessment based on temperature recovery capability. This invention solves the problems of difficulty in accurately identifying the overall temperature distribution and overheating risk inside the distribution box, and the tendency for false alarms or delayed warnings, caused by limited sensing nodes, susceptibility to temperature data interference, and fixed threshold judgments.

Description

Technical Field

[0001] This invention relates to the field of overheating warning technology for distribution boxes, and more specifically, to a method for overheating warning of distribution boxes with integrated wireless modules. Background Technology

[0002] As a crucial piece of equipment in power distribution systems, the distribution box's primary function is to distribute, control, and protect electrical energy. During long-term operation, areas within the distribution box, such as the circuit breaker terminals, busbar connections, cable joint crimping points, and contactor contacts, are prone to localized temperature increases due to the Joule heating effect generated by current flow and increased contact resistance. When the temperature continues to rise or forms localized hotspots, it not only accelerates the aging of insulation materials but may also lead to poor contact, equipment damage, or even electrical fires. Therefore, real-time monitoring of the internal temperature of the distribution box and timely warnings of overheating risks are of paramount importance for ensuring the safe and stable operation of power distribution equipment.

[0003] In existing technologies, common methods for monitoring the temperature of distribution boxes mainly include manual infrared thermography, fixed-point temperature sensor monitoring, and online temperature monitoring systems based on simple threshold judgment. Among them, manual infrared inspection can detect some abnormal heat points, but it has problems such as long detection cycle, poor real-time performance, and difficulty in covering hidden locations inside the distribution box; while fixed-point temperature sensor monitoring systems usually install temperature sensors in a few key locations and trigger overheat alarms by setting fixed temperature thresholds.

[0004] While this method enables online monitoring, the limited number of sensor nodes only allows for discrete temperature data points, making it difficult to reflect the overall temperature distribution inside the distribution box. Furthermore, it cannot effectively monitor areas without sensors. In addition, existing systems often employ simple filtering or direct sampling for temperature data processing, making them susceptible to electromagnetic interference, switching actions, or transient load changes, leading to fluctuations or deviations in temperature data and reducing monitoring accuracy. Moreover, traditional systems commonly use fixed temperature thresholds for early warning, failing to comprehensively consider factors such as equipment load changes, ambient temperature, and temperature trends, resulting in false alarms or delayed warnings in actual operation. Summary of the Invention

[0005] To overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a method for overheating warning of a distribution box with an integrated wireless module, which solves the problems in the prior art that make it difficult to accurately identify the overall temperature distribution and overheating risk inside the distribution box due to the limited number of sensing nodes, the susceptibility of temperature data to interference, and the fixed threshold judgment, and that are prone to false alarms or delayed warnings.

[0006] To achieve the above objectives, the present invention provides the following technical solution: Firstly, this application provides a method for overheating early warning of a distribution box with an integrated wireless module. The method includes: collecting temperature signals inside the distribution box and performing dynamic filtering and drift compensation processing on the temperature signals based on a joint determination of sliding window temperature offset trend and information entropy attenuation characteristics; dividing the processed signals into regions based on the spatial location of sensor nodes, and using the internal structural information of the distribution box to obtain complete temperature field data through structural constraint spatial interpolation to identify local hotspots; for local hotspot regions, constructing a thermodynamic model and determining an adaptive temperature threshold by combining temperature safety margin and temperature rise trend correction; comparing real-time temperature data with the adaptive temperature threshold, triggering an overheating early warning by using a dual determination mechanism of continuous over-threshold cumulative counting and duration, and conducting a risk assessment based on temperature recovery capability.

[0007] In one embodiment, the dynamic filtering and drift compensation processing based on the joint determination of sliding window temperature offset trend and information entropy decay characteristics includes: acquiring the original temperature data stream and performing outlier identification processing to obtain a first temperature data sequence with significant outliers removed; performing sliding window filtering processing on the first temperature data sequence and performing weighted summation according to preset weights to obtain a smoothed second temperature data sequence.

[0008] In one embodiment, the dynamic filtering and drift compensation processing based on the joint determination of temperature offset trend and information entropy decay characteristics within a sliding window further includes: establishing a temperature baseline model, calculating the temperature offset of the second temperature data sequence relative to the reference temperature within a continuous sliding window, and performing a consistency analysis on the temperature offset direction; generating candidate drift intervals when the temperature offsets of multiple consecutive sliding windows maintain the same sign and change direction; extracting the local temperature subsequences corresponding to the candidate drift intervals, calculating their information entropy feature values, and analyzing the information entropy changes of consecutive sliding windows; determining that there is an entropy decay trend when the information entropy changes of multiple consecutive sliding windows are all less than zero; determining that there is a systematic drift of the current temperature sensor if the temperature offset direction remains consistent and the information entropy feature value continues to decrease; generating drift compensation parameters based on the temperature offset and information entropy change amplitude corresponding to the systematic drift, and using the drift compensation parameters to correct the second temperature data sequence to obtain a third temperature data sequence.

[0009] In one embodiment, complete temperature field data is obtained through structural constraint spatial interpolation using the internal structural information of the distribution box to identify local hotspots. This includes: dividing the distribution box into regions based on the spatial location of the sensing nodes, and performing weighted summation on the third temperature data sequence within each region to obtain the equivalent temperature value of the region; establishing a thermal connectivity topology based on the internal structural information of the distribution box, determining effective adjacent regions based on the thermal connectivity topology and performing weighted interpolation to obtain the predicted temperature value; combining the equivalent temperature value of the region with the predicted temperature value to construct complete temperature field data; constructing a trend function for several consecutive complete temperature field data to predict the temperature value of the next sampling period, and combining this with the current temperature rise rate for joint judgment to identify local hotspots.

[0010] In one embodiment, joint judgment is performed to identify local hotspots, including: if the temperature rise rate is greater than a preset rate threshold for several consecutive cycles, or the predicted temperature value of the next sampling cycle is greater than a preset safe temperature threshold, or the difference between the current area's complete temperature field data and the average temperature of its adjacent areas is greater than a preset threshold, then abnormal marker data is generated for the area, and the corresponding area number and temperature parameters are output.

[0011] In one embodiment, determining effective adjacent regions based on the thermal connectivity topology graph and performing weighted interpolation includes: marking the reachable adjacent region set of each region node according to the thermal connectivity topology graph to form a structural constraint adjacency matrix; mapping the target prediction location without sensor nodes to the structural constraint adjacency matrix to obtain an effective region set; extracting the regional equivalent temperature value and spatial coordinates of each region in the effective region set, and calculating the effective distance data to the target prediction location; constructing a structural constraint weight set based on the distance data, and performing a weighted summation of the regional equivalent temperature values ​​of the effective region set to obtain the predicted temperature value of the target prediction location.

[0012] In one embodiment, for local hotspot areas, a thermodynamic model is constructed and combined with temperature safety margin and temperature rise trend correction to determine an adaptive temperature threshold. This includes: acquiring first data for the area corresponding to the local hotspot and establishing a continuous thermodynamic model for the region, wherein the first data includes complete temperature field data for each region, ambient temperature, and load rate of the corresponding region; discretizing the continuous thermodynamic model to establish a discrete temperature state space model for the region, and solving for the theoretical steady-state temperature of the system based on the current operating conditions; calculating the temperature safety margin based on the theoretical steady-state temperature to construct an energy margin function, and constructing a trend correction amount based on the temperature rise rate to correct the energy margin; and determining the adaptive temperature threshold based on the correction result.

[0013] In one embodiment, determining the adaptive temperature threshold includes: calculating the temperature safety margin between the current actual temperature and the theoretical limit temperature based on the theoretical steady-state temperature value, and constructing an energy margin function; constructing a trend correction amount based on the temperature rise rate, and constructing a comprehensive safety margin based on the energy margin function and the trend correction amount; and calculating the adaptive temperature threshold based on the comprehensive safety margin and the theoretical steady-state temperature.

[0014] In one embodiment, real-time temperature data is compared with an adaptive temperature threshold, and a dual determination mechanism of continuous over-threshold cumulative counting and duration is used to trigger an overheating warning. This includes: acquiring regional temperature data and comparing it with the adaptive temperature threshold to obtain the temperature status determination result for the current sampling period; based on the temperature status determination result, establishing an over-threshold duration counter for each monitoring area and updating the counter in each sampling period to form a periodic counting sequence; based on the periodic counting sequence, determining the current count value of the over-threshold duration counter in each sampling period and generating a stable anomaly determination identifier based on the determination result; based on the stable anomaly determination identifier, calculating the cumulative over-threshold duration for the area; comparing the cumulative over-threshold duration with a preset anomaly duration threshold, confirming that the temperature anomaly in the area is a stable overheating state based on the comparison result, and generating an overheating warning trigger signal.

[0015] In one embodiment, risk assessment is performed in conjunction with temperature recovery capability, including: acquiring overheating event record data; monitoring current regional temperature changes, and extracting temperature recovery process data segments when the temperature changes from an over-threshold state to a continuous decline and recovers to below the adaptive temperature threshold; calculating the temperature recovery time and temperature rise amplitude based on the data segments, constructing a temperature recovery capability evaluation parameter set, and then calculating the temperature recovery capability index; comparing the temperature recovery capability index with historical average recovery capability data to obtain a temperature recovery capability deviation parameter; constructing a comprehensive risk assessment index based on the temperature recovery capability index and its deviation parameter, classifying risk levels accordingly, and generating operational handling recommendations.

[0016] As can be seen from the above technical solutions, the embodiments of this application have the following advantages: Compared to conventional overheat monitoring solutions for distribution boxes that only employ fixed temperature thresholds, single-point temperature measurement, and simple over-temperature alarms, the innovation of the above technical solution lies in the following: It not only achieves adaptive identification and compensation for systematic sensor drift through sliding weighted filtering, information entropy attenuation analysis, and drift intensity quantification, reducing the accumulation of measurement errors during long-term operation, but also establishes a thermal connectivity constraint interpolation model based on the internal topology of the distribution box. This gives the temperature prediction of unmonitored areas a real physical meaning of heat propagation, solving the problem that traditional discrete-point temperature measurement cannot reflect the overall heat distribution. Simultaneously, it dynamically solves the theoretical steady-state temperature through a thermodynamic model and generates an adaptive temperature threshold by combining the temperature rise trend and safety margin. This allows the warning threshold to change in real time with load rate, ambient temperature, and heat dissipation status, avoiding false alarms and missed alarms caused by fixed thresholds. Furthermore, it introduces a continuous over-threshold accumulation judgment mechanism and a "temperature recovery capability" evaluation dimension, comprehensively analyzing the equipment's thermal state from both heating and cooling processes, enabling early identification of heat dissipation degradation, contact aging, and potential chronic thermal failures. Overall, this solution is significantly superior to conventional technical solutions in terms of temperature measurement accuracy, early hotspot identification, reliability of anomaly detection, and long-term operational risk assessment capabilities, and has stronger engineering applicability and intelligence. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic flowchart of an overheat warning method for a distribution box with an integrated wireless module, provided in an embodiment of this application.

[0019] Figure 2 The comparison curves of the original temperature data, filtered data, and drift compensation and correction data provided in the embodiments of this application are shown.

[0020] Figure 3 A comparison chart of prediction errors between the traditional interpolation algorithm and the structural constraint interpolation algorithm provided in the embodiments of this application. Detailed Implementation

[0021] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] Reference Figure 1 As shown, the present invention provides a method for overheating early warning of a power distribution box with an integrated wireless module, comprising the following steps: S1 collects real-time temperature signals at various points inside the distribution box by using wireless temperature sensing nodes in key areas.

[0023] In this embodiment, after the power distribution box is de-energized and its safety status is confirmed, the key components prone to heat generation are first identified and located based on the primary wiring diagram and actual internal layout of the distribution box. These key components include the circuit breaker inlet and outlet terminals, contactor contact positions, and branch circuit connection terminals. Then, based on the spatial dimensions, insulation distance requirements, and electrical safety specifications of each key component, wireless temperature sensing nodes of matching sizes are selected, and the installation locations are cleaned and insulated. Next, high-temperature resistant insulating fixing structures (such as insulating clips, heat-resistant pads, or special clamps) are used to tightly attach and fix the wireless temperature sensing nodes to the surface of the heat-generating components, ensuring full contact between the sensor's temperature measuring surface and the measured conductor to improve thermal response efficiency. Then, each wireless node is uniquely addressed and paired with the centralized receiving module, and the sampling period and data upload interval are set. Finally, after power is restored, a trial run test is conducted. By applying different current conditions, the stability, synchronization, and wireless transmission reliability of the temperature data collected by each node are verified, confirming that all temperature measuring nodes can continuously and accurately collect and transmit real-time temperature signals from various points inside the distribution box.

[0024] S2 performs preliminary filtering and calibration on the acquired temperature signal to eliminate transient interference and correct sensor drift, thereby reducing the cumulative effect of single-point sensor errors, including: The raw temperature signals output by each wireless temperature sensing node in multiple consecutive sampling periods are acquired and formed into a continuous raw temperature data stream in chronological order. Anomaly identification processing is performed based on the original temperature data stream. Through a preset reasonable temperature range, sampling points that exceed the reasonable temperature range are marked as anomalies, resulting in a first temperature data sequence with anomalies removed. The reasonable temperature range can be set according to the rated operating temperature range of the monitored equipment or environment, and in conjunction with the measurement upper and lower limit parameters in the temperature sensor product manual, as the range between the corresponding maximum and minimum allowable temperature values.

[0025] Based on the first temperature data sequence, a sliding window filtering process is performed to construct a time sliding window containing several consecutive sampling points, and weighted summation is performed to eliminate pulse fluctuations caused by transient electromagnetic interference or switching actions, so as to obtain a smoothed second temperature data sequence. The second temperature data sequence is calculated using the following formula: [Formula 1] Among them, weight satisfy: [Formula 2] In the formula, The temperature value of the i-th temperature sensing node at time t after sliding window filtering is the second temperature data sequence. The length of the sliding window. is the weight coefficient for the k-th historical sampling point, where k is the time index within the sliding window.

[0026] Furthermore, the preliminary filtering and calibration process also includes: A temperature baseline model is established. The baseline model selects a stable operating period when the load fluctuation rate of the distribution box is less than 5%, calculates the average temperature during this period as the reference temperature, obtains the average temperature of the second temperature data sequence, and calculates the temperature offset by combining the reference temperature. The specific formula for calculating the temperature offset is as follows: [Formula 3] In the formula, This is the temperature offset. This represents the average temperature of the second temperature data sequence within the current running window. This is the reference temperature.

[0027] Based on the temperature offset, a consistency analysis is performed on the temperature offset direction in multiple consecutive sliding windows: if the temperature offset in multiple consecutive sliding windows maintains the same sign and the temperature offset change direction between adjacent sliding windows is consistent, it is determined that the current temperature sequence has a continuous unidirectional offset trend, and a corresponding candidate drift interval is generated. For the candidate drift interval, the local temperature subsequence within the corresponding sliding window is extracted, and the local temperature subsequence is discretized and grouped according to the preset temperature interval to obtain multiple temperature state intervals. The number of sampling points in each temperature state interval is counted, and the corresponding state probability is calculated. The specific formula for calculating the state probability is as follows: [Formula 4] In the formula, Let be the state probability corresponding to the i-th temperature state interval. Let be the number of sampling points within the i-th temperature range. This represents the total number of temperature state intervals. This represents the total number of sample points across all temperature ranges.

[0028] The corresponding information entropy feature value is calculated based on the state probability to characterize the dynamic complexity of the current temperature sequence; The specific formula for calculating the information entropy feature value is as follows: [Formula 5] In the formula, This is the characteristic value of information entropy.

[0029] Entropy decay trend analysis is performed based on the aforementioned information entropy feature values: Calculate the change in information entropy between adjacent sliding windows; [Formula 6] In the formula, The change in information entropy Let be the information entropy feature value of the i-th sliding window. Let be the information entropy feature value of the (i-1)th sliding window.

[0030] If the change in information entropy corresponding to multiple consecutive sliding windows is less than zero, it is determined that the current temperature sequence has an entropy decay trend, and an entropy decay feature sequence is generated. The continuous unidirectional offset trend is correlated and matched with the entropy decay feature sequence. When the temperature offset direction is consistently consistent and the information entropy feature value continues to decrease, it is determined that the current temperature sensor has a systematic drift. The systematic drift is used to characterize the low randomness, low complexity, and continuous gradual change characteristics of temperature changes.

[0031] The drift intensity is calculated based on the temperature shift and information entropy change magnitude corresponding to the systematic drift. The specific formula for calculating the drift intensity is as follows: [Formula 7] In the formula, For drift strength, , These are the weighting coefficients, The number of sliding windows. This is the information entropy baseline value for the temperature sequence under normal operating conditions.

[0032] Drift compensation parameters are generated based on the drift intensity, and the drift compensation parameters are used to correct the second temperature data sequence to obtain the corrected third temperature data sequence.

[0033] The specific calculation formula for the correction process is as follows: [Formula 8] In the formula, This is the corrected third temperature data sequence.

[0034] like Figure 2 As shown, the original temperature data exhibits obvious random spikes and low-frequency drift, especially showing strong instability during switching operations and load fluctuations. After sliding window weighted filtering, high-frequency noise is significantly suppressed, and the smoothness of the temperature curve is improved. Furthermore, by introducing a drift compensation mechanism based on information entropy decay and trend consistency analysis, the long-term drift trend is effectively corrected. The corrected temperature sequence can more stably reflect the true thermal state, verifying the superiority of this embodiment in terms of anti-interference and long-term operational stability.

[0035] Table 1 shows the temperature standard deviation, mean square error, and drift bias for different treatment stages. The results indicate that this embodiment demonstrates higher stability in long-term drift suppression compared to the traditional moving average method.

[0036] Table 1 Comparison of Filtering Effects

[0037] It should be noted that, compared to conventional preliminary temperature processing methods that rely solely on fixed threshold filtering, traditional moving averages, or single-reference difference correction, the advantages of the above technical solution lie in constructing a multi-layered progressive processing framework: "anomaly removal—moving weighted filtering—baseline modeling—trend consistency determination—information entropy complexity analysis—drift intensity quantification—adaptive compensation correction." This framework not only effectively suppresses the impact of transient electromagnetic interference, switching operation pulses, and random noise on single-point measurements, but also further achieves adaptive calibration for differences in operating conditions by establishing a dynamic reference temperature model under low load fluctuation conditions. Simultaneously, it introduces a continuous window-based... The joint criterion of mouth offset consistency and information entropy decay characteristics identifies systematic drift from two dimensions: "numerical offset" and "sequence complexity degradation". Compared with traditional methods that rely solely on mean deviation or a single trend, this significantly reduces the risk of misjudging slow changes in real operating conditions as sensor drift and improves the reliability and robustness of drift identification. In addition, by quantitatively modeling drift intensity and driving the generation of compensation parameters, adaptive correction for drift of different degrees is achieved, enabling temperature data to maintain higher consistency and stability during long-term operation, thereby effectively suppressing the impact of single-point sensor error accumulation on system state assessment.

[0038] S3, based on the processed signal, divides the area according to the spatial location of the sensor nodes, and uses the internal structural information of the distribution box to obtain the predicted temperature value of the unmonitored area through structurally constrained spatial interpolation, thus constructing complete temperature field data to identify local hotspots, such as... Figure 3The figure shows a comparison of hotspot prediction errors between the traditional interpolation algorithm and the structural constraint interpolation algorithm of this embodiment. Experimental results show that the average prediction error of this embodiment is reduced by approximately 32.6% in unmonitored areas.

[0039] In this embodiment, the processed signal is used to divide the region according to the spatial location of the sensing node, and the predicted temperature value of the unmonitored area is obtained by structural constraint spatial interpolation using the internal structural information of the distribution box, thus forming complete temperature field data to identify local hotspots, including: Based on the internal structural coordinates of the distribution box, each temperature sensing node is spatially numbered, and a node spatial mapping table is established, in which each sensing node corresponds to a unique spatial coordinate. Based on the spatial mapping table, the Euclidean distance between any two sensing nodes is calculated. Sensing nodes whose Euclidean distance is less than the preset region division threshold are divided into the same region set, thereby forming several region node subsets. The preset region division threshold can be determined by taking the average value or 1 to 1.5 times the average value of the actual installation spacing of sensing nodes inside the distribution box or the typical node spacing in the design drawings as an empirical threshold. The Euclidean distance is specifically calculated using the following formula: [Formula 9] In the formula, Let be the Euclidean distance between sensor node i and sensor node j. Let be the spatial coordinates of sensor node i. Let be the spatial coordinates of sensor node j.

[0040] After completing the region division, for each sensor node in each region set, calculate the distance from each node to the geometric center of the region, and construct normalized weight coefficients according to the inverse distance method. The normalized weighting coefficient is calculated using the following formula: [Formula 10] In the formula, Let be the normalized weighting coefficient of the i-th temperature sensing node within its region. Let n represent the distance from the i-th temperature sensing node to the geometric center of the region, where n is the number of temperature sensing nodes contained in the region. It is the sum of the reciprocals of the distances to all nodes within the region.

[0041] Based on normalized weighting coefficients, the third temperature data sequence of each node in this region is... Weighted summation is performed to obtain the equivalent temperature value of the region, so as to reduce the impact of single-node error on region judgment; The equivalent temperature value of the region is calculated using the following formula: [Formula 11] In the formula, This is the equivalent temperature value for the region, a comprehensive temperature value used to characterize the overall thermal state of the region. is the normalized weighting coefficient of the i-th temperature sensing node within its region.

[0042] A thermal connectivity topology map is established based on the internal structural information of the distribution box. The effective adjacent areas are determined according to the thermal connectivity topology map and weighted interpolation is performed to obtain the predicted temperature value. The predicted temperature value is combined with the equivalent temperature values ​​of all known regions to form complete temperature field data. Complete temperature field data is stored in chronological order to form regional temperature history sequence data, and the current temperature rise rate is calculated in each sampling period to reflect the intensity of temperature change in the region. The specific formula for calculating the temperature rise rate is as follows: [Formula 12] In the formula, For the rate of temperature rise, This provides complete temperature field data for the current sampling period. This provides complete temperature field data from the previous sampling period. This represents the sampling time interval.

[0043] Linear trend calculation is performed on complete temperature field data for several consecutive sampling periods. Least squares linear fitting is used to obtain the trend function, and the predicted temperature value for the next sampling period is calculated by extrapolation based on the trend function. The trend function is calculated using the following formula: [Formula 13] The predicted temperature value for the next sampling period is calculated using the following formula: [Formula 14] In the formula, To fit the temperature value over time variable t, The slope of the trend function represents the rate of temperature change over time. The intercept term of the trend function. This is the predicted temperature value for the next sampling period.

[0044] By combining the current complete temperature field data, the predicted temperature value for the next sampling period, and the temperature rise rate, local hotspots can be identified.

[0045] Furthermore, the joint judgment to identify local hotspots includes: If the temperature rise rate exceeds the preset rate threshold for several consecutive cycles (which can be set based on the statistical results of historical temperature rise rate data of the power distribution equipment under normal steady-state operation conditions (such as the 95th percentile or the mean plus three times the standard deviation)), or the predicted temperature value of the next sampling cycle exceeds the preset safe temperature threshold (which can be directly determined based on the rated maximum allowable operating temperature of each component or conductor in the distribution box, the temperature rise limit specified by national or industry standards, and the upper limit temperature parameter in the manufacturer's technical specifications, and the minimum safe value is taken as the system threshold), or the difference between the current area's complete temperature field data and the average temperature of its adjacent areas exceeds the set temperature difference threshold (which can be determined based on the statistical distribution characteristics of the historical temperature difference between adjacent areas in the same distribution box under normal operating conditions (such as the mean plus two or three times the standard deviation) or the allowable deviation range of thermal equilibrium commonly found in engineering experience), then abnormal marker data will be generated for that area, and the corresponding area number and temperature parameters will be output.

[0046] It should be noted that by spatially numbering and dividing the processed temperature data into regions, adjacent nodes are constructed into a set of regions, and the equivalent temperature value of the region is calculated using distance reciprocal normalization weighting. This reduces the impact of single-point sensing errors and random fluctuations on the overall judgment from the source. At the same time, a structural constraint interpolation algorithm is used to predict the temperature of unmonitored areas under physically reachable paths, constructing complete temperature field data and expanding the monitoring results from discrete points to a continuous spatial distribution. On this basis, the intensity of temperature change and future trend are dually characterized by temperature rise rate calculation and least squares trend extrapolation. The current temperature, predicted temperature, and regional temperature difference are jointly judged to avoid relying on a single threshold to trigger false alarms or missed alarms. Thus, while ensuring controllable computational complexity, the accuracy, advance warning, and stability of local hotspot identification are improved, enhancing the engineering reliability and practicality of the overheating early warning system for distribution boxes.

[0047] Furthermore, a thermal connectivity topology map is established based on the internal structural information of the distribution box. Effective adjacent regions are determined according to the thermal connectivity topology map, and weighted interpolation is performed to obtain the predicted temperature value, including: Obtain the internal structure information of the distribution box, including the position of the partition, the position of the conductive busbar, the position of the air passage and the position of the shell boundary. Based on the internal structure information, establish an internal spatial topology diagram, take the geometric center of each divided area as a node, and establish connecting edges between areas with actual heat conduction or convection channels to form a thermal connectivity topology diagram. Based on the thermal connectivity topology, each region node is marked with its set of reachable adjacent regions. If two regions are blocked by a metal partition, insulation barrier or closed structure and there is no effective heat conduction or air flow path, the connection between the two regions is canceled in the topology, thereby forming a structural constraint adjacency matrix that only contains the actual heat propagation path. After establishing the structural constraint adjacency matrix, the predicted location of the target without sensor nodes is determined, and according to its spatial location attribution, the predicted location of the target is mapped to the structural unit region where it is located or the set of adjacent regions that are thermally connected to it, so as to obtain the effective region set for interpolation calculation. Based on the three-dimensional structural division of the distribution box, the internal space is divided into several structural unit regions according to partitions, busbars, air channels, and shell boundaries. Each structural unit region is assigned a unique number and spatial boundary range. Then, the three-dimensional coordinates of the target prediction location without sensor nodes are obtained. By determining whether the coordinates fall within the spatial boundary range of a certain structural unit region, the structural unit region number to which it belongs is determined. If the target prediction location is located at the intersection of two or more structural units, its primary region is determined according to the shortest spatial distance principle or the thermal connectivity priority principle. After determining the structural unit region, combined with the aforementioned structural constraint adjacency matrix, all adjacent region nodes that have thermal connectivity with the region in the topology graph are retrieved to form a set of regions with actual heat propagation paths to the target prediction location. This set of regions is determined as the effective set of regions to participate in subsequent interpolation calculations, thereby ensuring that temperature prediction is only calculated within the actual heat conduction or convection reach range, improving the physical rationality and structural consistency of spatial prediction.

[0048] The equivalent temperature value and corresponding spatial coordinates of each region are extracted within the set of effective regions. The spatial distance between the predicted target location and the center of the effective region is calculated to obtain effective distance data, which is obtained by Euclidean distance calculation. Based on the effective area distance data, construct the structural constraint weight coefficients. When there is no hot-connected path between a certain region and the target prediction location, its structural constraint weight coefficient is set to zero. Only regions with hot-connected paths are weighted using the inverse distance method and then normalized to obtain the desired result. The set of structural constraint weights; Based on the set of structural constraint weights, the equivalent temperature values ​​of the effective region are weighted and summed to obtain the predicted temperature value of the target prediction location in the current sampling period.

[0049] It should be noted that by introducing topological constraints on the internal structure of the distribution box before spatial interpolation calculation, the actual physical structures such as partitions, busbars, air channels, and shell boundaries are included in the heat propagation path determination process. This ensures that temperature prediction is only performed between areas where heat conduction or convection channels actually exist, thus avoiding the problem of unreasonable calculations across insulation structures or enclosed spaces in traditional inverse distance weighted interpolation, and reducing the ineffective spatial diffusion of errors. At the same time, through structural unit attribution determination and adjacency matrix screening mechanisms, the effective set of regions participating in the calculation is limited, reducing the interference of irrelevant regions on the prediction results, improving the pertinence and physical consistency of local hotspot prediction, and enhancing the stability and engineering applicability of temperature distribution reconstruction while ensuring controllable computational complexity. This overall improves the accuracy and reliability of temperature prediction in unmonitored areas inside the distribution box.

[0050] S4. For the identified local hotspot regions, an independent thermodynamic model is constructed to solve for the theoretical steady-state temperature in the region. Combined with the temperature safety margin and temperature rise trend correction, the adaptive temperature threshold of the local hotspot region is determined.

[0051] In this embodiment, by constructing a thermodynamic model of a local hotspot region and solving for the theoretical steady-state temperature, and combining this with temperature safety margin and temperature rise trend correction, an adaptive temperature threshold is determined, including: First data of the region corresponding to the local hotspot is obtained. The first data includes complete temperature field data of each region, ambient temperature and load rate of the corresponding region. The region is regarded as a lumped parameter thermal system and a continuous thermodynamic model of the region is established based on the principle of thermal energy conservation. The specific calculation formula for the thermodynamic continuity model is as follows: [Formula 15] In the formula, This is the equivalent heat capacity of the region, used to characterize the amount of heat absorbed or released required for a temperature change in that region. For complete temperature field data, The heat generated by the load. For load rate, The equivalent heat dissipation coefficient, The ambient temperature.

[0052] Among them, the thermodynamic continuous model is a mathematical model that describes the process of heat generation, transfer and dissipation inside the equipment through continuous-time differential equations based on thermodynamic mechanisms such as heat conduction, heat convection and heat capacity, thereby characterizing the continuous change of system temperature over time.

[0053] After establishing the aforementioned thermodynamic continuity model, based on the Joule heating characteristics of conductors, the heating power is expressed as a function of the load factor, specifically: [Formula 16] In the formula, The equivalent heating coefficient is used to describe the change in heating power caused by a change in unit load.

[0054] The thermodynamic continuous model is discretized, and the sampling period is set to be... A discrete model of the regional temperature state space is established by using the forward Euler discretization method; The specific calculation formula for the temperature state space discrete model is as follows: [Formula 17] in, , , In the formula, The retention ratio of the current temperature at the next moment is the temperature inertia coefficient. The load heating coefficient represents the degree of temperature increase caused by the load's thermal power per unit time. The ambient temperature coupling coefficient represents the degree to which the equipment temperature approaches the ambient temperature within a time step. The load heat power conversion factor represents the proportionality of heat generated by the square of the load. The equivalent heat exchange coefficient represents the heat dissipation capacity generated per unit temperature difference between the equipment and the environment.

[0055] Among them, the temperature state-space discrete model is a mathematical expression that discretizes the continuous thermodynamic model under discrete sampling periods, uses temperature as the state variable, and describes the temperature state evolution relationship between adjacent sampling times through state transition equations and input terms.

[0056] After establishing the temperature state space discrete model, the theoretical steady-state temperature of the system is solved according to the current operating conditions. When the system reaches thermal equilibrium, it satisfies the following conditions: , which means that the rate of change of the system's state variable T with respect to time is zero, that is, the equipment temperature no longer changes with time, the system has reached thermal equilibrium, and thus the theoretical steady-state temperature value of the region is obtained; The theoretical steady-state temperature value is calculated using the following formula: [Formula 18] In the formula, This is the theoretical steady-state temperature value, representing the theoretically achievable thermal equilibrium limit temperature for this region under the current load rate and ambient temperature conditions. The equivalent heat dissipation coefficient, The equivalent heating coefficient, The ambient temperature.

[0057] Based on the theoretical steady-state temperature, the temperature safety margin is calculated to construct an energy margin function. A trend correction amount is constructed according to the temperature rise rate to correct the energy margin. The adaptive temperature threshold is determined based on the correction result.

[0058] Furthermore, an energy margin function is constructed based on the theoretical steady-state temperature to calculate the temperature safety margin, and a trend correction is constructed according to the temperature rise rate to correct the energy margin. An adaptive temperature threshold is then determined based on the correction result, including: Based on the theoretical steady-state temperature value, calculate the temperature safety margin between the current actual temperature and the theoretical limit temperature, and construct the energy margin function; The energy margin function is calculated using the following formula: [Formula 19] In the formula, This represents the remaining temperature safety margin of the current system before reaching the theoretical thermal limit. This provides complete temperature field data.

[0059] The energy margin function is a function used to characterize the remaining heat capacity or thermal safety margin between the current thermal energy state of the system and the safe operating limit. It quantifies the thermal safety margin of the equipment by calculating the difference between the actual thermal energy of the system and the maximum allowable thermal energy.

[0060] The trend correction amount is constructed based on the rate of temperature rise; The specific formula for calculating the trend correction amount is as follows: [Formula 20] In the formula, This is the trend correction amount. This is a trend adjustment coefficient used to describe the degree of influence of the temperature rise rate on the safety margin. This represents the rate of temperature rise.

[0061] Based on the trend correction amount, a comprehensive safety margin is constructed according to the energy margin function and the trend correction amount; The comprehensive safety margin is calculated using the following formula: [Formula 21] In the formula, This is to account for the actual available safety margin after considering the upward trend in temperature.

[0062] Based on the comprehensive safety margin, the adaptive temperature threshold is calculated in conjunction with the theoretical steady-state temperature. The adaptive temperature threshold is calculated using the following formula: [Formula 22] In the formula, The adaptive temperature threshold represents the dynamic safety temperature boundary that the system is allowed under the current thermodynamic state and temperature change trend.

[0063] It should be noted that by establishing a continuous thermodynamic model for local hotspot areas and discretizing the steady-state temperature, the temperature threshold no longer depends on fixed empirical values, but is matched with the actual load rate, ambient temperature, and heat dissipation capacity. At the same time, an energy margin function is introduced to quantify the safe space between the current temperature and the thermal limit, and a trend correction quantity is constructed in combination with the temperature rise rate to dynamically compensate for the safety margin. This results in an adaptive temperature threshold that changes in real time with the operating status. Compared with the traditional static threshold method, it can more accurately reflect the actual thermal balance state and temperature rise trend of the equipment, effectively avoid false alarms or missed alarms caused by changes in operating conditions, and significantly improve the accuracy of local hotspot identification and the robustness of the early warning system.

[0064] S5 compares real-time temperature data with adaptive temperature thresholds and adopts a dual judgment mechanism of continuous over-threshold cumulative counting and duration. It only triggers an overheating warning signal when the area temperature continuously exceeds the threshold and the cumulative duration exceeds the preset threshold. The signal is then transmitted to the centralized monitoring system via a wireless module and combined with temperature recovery capability for risk assessment.

[0065] In this embodiment, triggering the overheating warning signal includes: The temperature data of each monitoring area within the current sampling period is obtained, and the adaptive temperature threshold of the corresponding area is obtained. The temperature data of the area is compared with the adaptive temperature threshold. When the temperature data of the area is greater than or equal to the adaptive temperature threshold, an over-threshold status mark is generated; otherwise, a normal status mark is generated, thereby obtaining the temperature status determination result of the current sampling period. Based on the temperature state determination results, an over-threshold continuous counter is established for each monitoring area, and the counter is updated in each sampling period. When it is determined to be an over-threshold state, the counter is incremented, and when it is determined to be a normal state, the counter is reset to zero, thereby forming a periodic counting sequence that reflects the continuous over-threshold process of temperature. Based on the periodic counting sequence, the current count value is compared with a preset threshold for exceeding the threshold. When the count value exceeds the threshold, the region is determined to have entered a stable threshold state, and a stable anomaly determination flag is generated. The threshold for determining the over-threshold can be set according to the temperature sampling period and the allowable duration of instantaneous fluctuations under normal operating conditions of the distribution box, so that short-term fluctuations will not trigger stable over-threshold determination within several consecutive sampling periods.

[0066] Based on the stable anomaly determination identifier, the cumulative over-threshold duration of the corresponding region is calculated. The cumulative over-threshold duration is obtained by multiplying the number of consecutive over-threshold sampling cycles by the duration of a single sampling cycle, so as to obtain a time parameter characterizing the degree of anomaly duration. The cumulative over-threshold duration is compared with the abnormal duration threshold. When the cumulative over-threshold duration exceeds the abnormal duration threshold, the temperature abnormality in the area is confirmed to be a stable overheating state, and an overheating warning trigger signal is generated.

[0067] The abnormal duration threshold can be set according to the typical temperature rise duration of the distribution box under normal load fluctuation conditions, so that it is only judged as a stable overheating state when the abnormal temperature duration exceeds the allowable range of normal operating conditions.

[0068] It should be noted that by comparing real-time temperature data with an adaptive temperature threshold, and introducing a continuous sampling cycle counting and cumulative duration determination mechanism when the temperature exceeds the threshold, the overheat warning signal is triggered only after multi-cycle stability confirmation of the temperature exceeding the threshold. This effectively avoids false alarms caused by short-term temperature spikes due to switching actions, electromagnetic interference, or instantaneous load fluctuations. At the same time, by comprehensively judging the number of consecutive over-threshold cycles and the cumulative duration, the system can more accurately distinguish between transient temperature fluctuations and real continuous overheating processes, thereby improving the stability and reliability of the distribution box overheat warning judgment and enhancing the overall warning mechanism's adaptability to actual operating conditions.

[0069] Furthermore, data is transmitted wirelessly to a centralized monitoring system, and a risk assessment is conducted based on temperature recovery capabilities, including: Based on the overheating warning trigger signal, overheating event record data is generated. The thermal event record data includes the over-threshold start time, peak temperature, duration information, area number, and timestamp information for the current area. The temperature changes in the current area are continuously monitored. When the temperature changes from an over-threshold state to a continuous decrease and recovers to below the adaptive temperature threshold, it is determined to be a temperature recovery process. The corresponding temperature recovery process data segment is extracted based on the overheating event record data. The temperature recovery process data segment refers to the continuous temperature data and its time series set from the moment the temperature in the region reaches its peak after the overheating event is confirmed until the temperature continues to drop and recovers to below the adaptive temperature threshold.

[0070] Based on the data fragments of the temperature recovery process, the temperature recovery time and the temperature rise amplitude are calculated. The temperature recovery time is the length of time it takes for the temperature to drop from the peak value to below the recovery threshold, and the temperature rise amplitude is the difference between the peak temperature and the over-threshold starting temperature, thereby constructing a set of evaluation parameters for temperature recovery capability. Based on the temperature recovery capability evaluation parameter set, the temperature recovery capability index is calculated to characterize the heat dissipation recovery capability of the corresponding area after overheating. The specific calculation formula for the temperature recovery capability index is as follows:

[0071] In the formula, As an indicator of temperature recovery ability, Temperature recovery time This represents the temperature rise.

[0072] The temperature recovery capability index and the corresponding overheating event record data are encoded and encapsulated to generate a temperature recovery capability assessment data frame, which is then sent to the centralized monitoring system through the wireless communication module inside the distribution box. After receiving the temperature recovery capability assessment data frame, the centralized monitoring system parses it to obtain the temperature recovery capability index of the corresponding area, and retrieves historical recovery capability data for comparative analysis to obtain the temperature recovery capability deviation parameter. The temperature recovery capability deviation parameter is calculated by the difference between the current temperature recovery capability index and the historical average recovery capability index.

[0073] Based on the temperature recovery capacity index and its deviation parameter, a comprehensive risk assessment index is constructed, and the risk level of the region is classified according to the comprehensive risk assessment index. The comprehensive risk assessment indicators are calculated using the following formulas:

[0074] In the formula, As a comprehensive risk assessment indicator, As an indicator of temperature recovery ability, This is a parameter representing the deviation of temperature recovery capability. , This is the risk assessment weighting coefficient.

[0075] Based on the comprehensive risk assessment index, the risk level of the abnormal area is divided. When the comprehensive risk assessment index reaches different preset level ranges, it is marked as low risk warning, medium risk warning or high risk warning respectively, so as to obtain the risk level assessment result of the corresponding area. Based on the risk level assessment results and the structural location and historical anomaly patterns of the corresponding area, the centralized monitoring system automatically generates corresponding operational handling suggestions. These suggestions include recommendations to reduce load, conduct key inspections, strengthen heat dissipation measures, or conduct emergency power outages for maintenance. The risk level assessment results and operational handling suggestions are then displayed or pushed to the operation and maintenance management terminal, thereby enabling real-time early warning and decision support for the risk of overheating in the distribution box.

[0076] It should be noted that, compared to conventional technical solutions in this field that rely solely on current temperature values, fixed temperature thresholds, or instantaneous temperature rise rates for risk assessment, this solution innovatively introduces "temperature recovery capability" as a risk assessment dimension. After an overheating event, it further analyzes the dynamic cooling process of the equipment recovering from the peak temperature to the safe temperature range. By constructing a temperature recovery capability index through temperature recovery time and temperature rise amplitude, it reflects the actual heat dissipation recovery performance of the area from the perspective of heat release efficiency. At the same time, it calculates the recovery capability deviation by combining historical recovery capability data to identify the degradation trend of the equipment's heat dissipation performance over operating time. This innovation not only breaks through the traditional solution's single judgment method that only focuses on the "heating process," but also reverses the assessment from the "cooling recovery process" to evaluate potential hidden dangers such as internal heat accumulation, obstructed ventilation, contact aging, or decreased heat dissipation capacity. This improves the long-term thermal fault identification capability and the accuracy of risk assessment, and enhances the system's early warning capability for early chronic thermal anomalies.

[0077] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0078] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0079] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0080] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. 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.

Claims

1. A method for overheating early warning of a distribution box integrating a wireless module, characterized in that, include: The temperature signal inside the distribution box is collected, and the temperature signal is subjected to dynamic filtering and drift compensation processing based on the joint determination of sliding window temperature offset trend and information entropy decay characteristics. Based on the processed signal, the area is divided by the spatial location of the sensor node, and the complete temperature field data is obtained by structural constraint spatial interpolation using the internal structural information of the distribution box in order to identify local hot spots; For local hotspot areas, a thermodynamic model is constructed and combined with temperature safety margin and temperature rise trend correction to determine an adaptive temperature threshold; Real-time temperature data is compared with adaptive temperature thresholds, and a dual judgment mechanism of continuous over-threshold cumulative counting and duration is used to trigger overheating warnings, and risk assessment is carried out in combination with temperature recovery capabilities.

2. The method for overheating early warning of a distribution box with an integrated wireless module according to claim 1, characterized in that, The dynamic filtering and drift compensation processing based on the joint determination of sliding window temperature offset trend and information entropy decay characteristics includes: The raw temperature data stream is acquired and outlier identification is performed to obtain the first temperature data sequence after removing significant outliers. A sliding window filter is applied to the first temperature data sequence, and then a weighted sum is performed according to preset weights to obtain a smoothed second temperature data sequence.

3. The method for overheating early warning of a distribution box with an integrated wireless module according to claim 2, characterized in that, The dynamic filtering and drift compensation processing based on the joint determination of sliding window temperature offset trend and information entropy decay characteristics also includes: A temperature baseline model is established, the temperature offset of the second temperature data sequence relative to the reference temperature within a continuous sliding window is calculated, and the consistency of the temperature offset direction is analyzed. When the temperature offsets of multiple consecutive sliding windows maintain the same sign and the direction of change is consistent, candidate drift intervals are generated. Extract the local temperature subsequence corresponding to the candidate drift interval, calculate its information entropy feature value, and analyze the information entropy change of consecutive sliding windows. When the information entropy change of multiple consecutive sliding windows is less than zero, it is determined that there is an entropy decay trend. If the temperature shift direction remains consistent and the information entropy characteristic value continues to decrease, it is determined that the current temperature sensor has a systematic drift. Based on the temperature offset and information entropy change magnitude corresponding to the systematic drift, drift compensation parameters are generated. The drift compensation parameters are then used to correct the second temperature data sequence to obtain the third temperature data sequence.

4. The method for overheating early warning of a distribution box with an integrated wireless module according to claim 1, characterized in that, The method of using the internal structural information of the distribution box to obtain complete temperature field data through structurally constrained spatial interpolation to identify local hot spots includes: The regions are divided into sets based on the spatial location of the sensor nodes, and the third temperature data sequence in each region is weighted and summed to obtain the equivalent temperature value of the region. A thermal connectivity topology map is established based on the internal structural information of the distribution box. The effective adjacent areas are determined according to the thermal connectivity topology map and weighted interpolation is performed to obtain the predicted temperature value. By combining the equivalent temperature value of the region with the predicted temperature value, a complete temperature field data is constructed. A trend function is constructed from several consecutive complete temperature field data to predict the temperature value of the next sampling period. Combined with the current temperature rise rate, a joint judgment is made to identify local hot spots.

5. The method for overheating early warning of a distribution box with an integrated wireless module according to claim 4, characterized in that, The joint judgment to identify local hotspots includes: If the temperature rise rate exceeds the preset rate threshold for several consecutive cycles, or the predicted temperature value of the next sampling cycle exceeds the preset safe temperature threshold, or the difference between the current region's complete temperature field data and the average temperature of its adjacent regions exceeds the preset threshold, then abnormal marker data will be generated for that region, and the corresponding region number and temperature parameters will be output.

6. The method for overheating early warning of a distribution box with an integrated wireless module according to claim 4, characterized in that, The step of determining the effective adjacent regions based on the hot connectivity topology graph and performing weighted interpolation includes: Based on the hot-connected topology graph, mark the set of reachable adjacent regions for each region node to form a structural constraint adjacency matrix; The predicted location of the target without sensor nodes is mapped to the structural constraint adjacency matrix to obtain the effective region set; Extract the equivalent temperature value and spatial coordinates of each region in the effective region set, and calculate the effective distance data to the predicted target location. Based on the distance data, a set of structural constraint weights is constructed, and the equivalent temperature values ​​of the effective region set are weighted and summed to obtain the predicted temperature value of the target prediction location.

7. The method for overheating early warning of a distribution box with an integrated wireless module according to claim 1, characterized in that, For local hotspot areas, a thermodynamic model is constructed and, combined with temperature safety margin and temperature rise trend corrections, an adaptive temperature threshold is determined, including: First data of the region corresponding to the local hotspot is obtained, and a continuous thermodynamic model of the region is established. The first data includes complete temperature field data of each region, ambient temperature and load rate of the corresponding region. The thermodynamic continuous model is discretized to establish a regional temperature state space discrete model, and the theoretical steady-state temperature of the system is solved according to the current operating conditions. Based on the theoretical steady-state temperature, the temperature safety margin is calculated to construct an energy margin function. A trend correction amount is constructed according to the temperature rise rate to correct the energy margin. The adaptive temperature threshold is determined based on the correction result.

8. The method for overheating early warning of a distribution box with an integrated wireless module according to claim 7, characterized in that, The determination of the adaptive temperature threshold includes: Based on the theoretical steady-state temperature value, calculate the temperature safety margin between the current actual temperature and the theoretical limit temperature, and construct the energy margin function; A trend correction is constructed based on the rate of temperature rise, and a comprehensive safety margin is constructed based on the energy margin function and the trend correction. Based on the comprehensive safety margin, the adaptive temperature threshold is calculated in conjunction with the theoretical steady-state temperature.

9. The method for overheating early warning of a distribution box with an integrated wireless module according to claim 1, characterized in that, The process of comparing real-time temperature data with an adaptive temperature threshold and triggering an overheating warning using a dual determination mechanism of continuous over-threshold accumulation and duration includes: The system acquires regional temperature data and compares it with an adaptive temperature threshold to obtain the temperature status determination result for the current sampling period. Based on the temperature status determination results, an over-threshold continuous counter is established for each monitoring area, and the counter is updated in each sampling period to form a periodic counting sequence. Based on the periodic counting sequence, the current count value of the over-threshold continuous counter is judged in each sampling period, and a stable anomaly judgment label is generated according to the judgment result. Based on the stable anomaly identification marker, calculate the cumulative over-threshold duration in this region; The cumulative over-threshold duration is compared with the preset abnormal duration threshold. Based on the comparison result, it is confirmed that the temperature abnormality in the area is a stable overheating state, and an overheating warning trigger signal is generated.

10. The method for overheating early warning of a distribution box with an integrated wireless module according to claim 1, characterized in that, The risk assessment based on temperature recovery capability includes: Acquire overheat event log data; Monitor the temperature change in the current area, and when the temperature changes from an over-threshold state to a continuous decrease and recovers to below the adaptive temperature threshold, extract the temperature recovery process data segment; Based on the data fragments, the temperature recovery time and temperature rise amplitude are calculated, a set of temperature recovery capability evaluation parameters is constructed, and then the temperature recovery capability index is calculated. The temperature recovery capability index is compared with historical average recovery capability data to obtain the temperature recovery capability deviation parameter. Based on the temperature recovery capability index and its deviation parameter, a comprehensive risk assessment index is constructed, and risk levels are classified accordingly, and operational and disposal recommendations are generated.

Images