Power distribution box remote inspection management method and system based on internet of things

Abstract

The application discloses a power distribution box remote inspection management method and system based on an Internet of Things, and relates to the technical field of power distribution equipment operation and maintenance management. In view of the problem that remote inspection defaults to adopt sensing data, and after sensing link failure, back transmission of smooth false data forms virtual inspection, the method carries verification incentive configuration in the inspection instruction, executes switching of incentive objects selected from existing controllable loads in the power distribution box by the intelligent gateway, and synchronously collects the output of each sensor in the verification observation window; item-by-item comparison is performed according to three types of expected response modes of change, slow change and invariance of the incentive response verification table, link credibility and link abnormality are distinguished; only when verification is passed, the inspection data is adopted, an inspection work order is dispatched for an abnormal channel, and previous data is marked for rechecking according to a tracing period, and if the abnormality exceeds an upper limit, manual inspection is performed, remote alarm is temporarily stopped, and the inspection interval is adaptively adjusted according to the verification passing condition. The application can verify the authenticity of the sensing link without adding hardware, and is suitable for remote inspection of the power distribution box.

Description

Technical Field

[0001] This invention relates to the field of power distribution equipment operation and maintenance management technology, and more specifically, to a remote inspection and management method for power distribution boxes based on the Internet of Things, as well as a system, electronic equipment and computer-readable storage medium for implementing the method. Background Technology

[0002] The number of low-voltage distribution boxes in industrial parks, buildings, and rural areas is large and scattered. According to the requirements of the "Distribution Network Operation Regulations", periodic inspections are required. Traditional manual inspections are time-consuming, labor-intensive, and subject to weather and traffic constraints. Therefore, installing current, temperature, door magnetic sensors, and other sensors and smart gateways on distribution boxes, and having the platform collect data remotely and periodically to replace manual on-site inspections, has become a common practice in the industry. The coverage of remote inspections has expanded rapidly in recent years.

[0003] However, remote inspections are based on a premise that is rarely scrutinized: the sensor data itself is reliable. Real-world failures often stem from this very premise. A temperature probe, after its adhesive has come off, is left suspended in the air and reports a constant room temperature reading; a loose acquisition terminal causes the current channel to consistently output the last valid value; a sensor power supply module malfunctions and its output slowly drifts—the common thread in these failures is that the transmitted data remains "stable and normal." The platform judges everything as normal based on the data, the inspection records are neat and accurate, but the true condition of the distribution box is unknown. The maintenance industry calls this a "false inspection": the inspection is completed in form, the monitor itself is faulty, and the monitored hidden dangers are masked by "normal data." In one industrial park, an overload and burn-out accident occurred in a distribution box. Subsequent investigation revealed that the box's current channel had been transmitting almost unchanged values ​​for several weeks due to terminal oxidation, while the platform's remote inspection records remained normal during this period—this is a typical consequence of a false inspection.

[0004] Current technologies are largely ineffective against false inspections. Manual inspections conducted according to regulations such as the "Typical Fire Safety Regulations for Power Equipment" can detect these problems, but the process is time-consuming, taking years and proving insufficient to address immediate needs. One type of remote solution enhances data processing analysis during inspections, introducing edge computing and time-series models for initial anomaly screening and risk prediction. However, the model identifies anomalies in the data, while the constant or slowly drifting data output by sensor link failures are statistically the most "stable." The model not only fails to alarm but also treats false data as reliable input for prediction, making the amplification of false data more subtle with increasing complexity. Another approach focuses on the sensor itself, adding a verification signal source and circuitry inside to periodically inject known signals to verify the sensor response. This type of self-calibrating sensor can detect sensor failures, but the verification signal stops within the sensor itself; the subsequent acquisition lines, terminals, acquisition circuits, and communication links are not within the verification scope. Loose terminals and cable issues are the most common problems observed in field statistics. A more practical obstacle is that dedicated verification hardware means replacing existing sensors, which is difficult for large and widely distributed distribution boxes to handle.

[0005] Through repeated comparisons and verifications on the front lines of operations and maintenance, the inventors discovered that the overlooked resource actually lies within the distribution box: the distribution box already contains existing loads controlled by the gateway, such as dehumidifiers, cooling fans, and internal lighting. Connecting or disconnecting any of these will cause predictable changes in the direction and magnitude of physical quantities such as current and temperature within the box—essentially, the distribution box itself contains a zero-cost excitation source. Each switching verifies the true response capability of the entire sensing link, from the sensing element to the platform's data entry. Existing remote inspection systems have never used this resource for data reliability verification; inspections are always one-way readings, lacking an active "question and answer" process.

[0006] Therefore, there is a need for a remote inspection management method that does not require any additional verification hardware, utilizes the existing controllable load in the distribution box to verify the authenticity of the sensor link before each round of inspection, and determines the data acceptance and inspection rhythm based on the verification results. Summary of the Invention

[0007] In view of this, regarding the reliability of remote inspection data, this invention is based on the principle of verification before trust, that is, each round of inspection first verifies the real response capability of the sensor link with a small and controllable stimulus, and only after the verification is passed can the data be trusted. This invention proposes a remote inspection management method for distribution boxes based on the Internet of Things, as well as the system, electronic equipment and computer-readable storage medium for implementing the method. The aim is to plug the loophole of false inspections in a way that requires zero additional hardware, so that every conclusion of remote inspection is based on verified data.

[0008] According to one aspect of the present invention, a remote inspection and management method for a distribution box based on the Internet of Things (IoT) is provided. The method includes: step S10, whereby an inspection management platform sends an inspection instruction to a smart gateway in the distribution box via an IoT communication link according to an inspection plan. The inspection instruction carries a verification incentive configuration, which specifies the incentive object selected from existing controllable loads in the distribution box and the duration of the incentive action; step S20, whereby the smart gateway performs switching on the incentive object according to the verification incentive configuration, and synchronously collects the outputs of each sensor in the distribution box within the verification observation window before and after the switching, wherein each sensor includes at least an outgoing current sensor, a box temperature sensor, and a box door status sensor; and step S30, whereby the smart gateway compares the outputs of each sensor in the verification observation window with a pre-stored incentive response verification table item by item, wherein the incentive response verification table records the pre-stored values ​​of each sensor according to the incentive object. The expected response mode is divided into three categories: expected change, expected gradual change, and expected no change. Sensors whose outputs match are considered reliable links, while those that do not are considered abnormal links. In step S40, when all sensors are considered reliable links, the inspection data of this round is accepted and an inspection conclusion is generated. When some sensors are abnormal, only the data of the reliable channel is accepted to generate an inspection conclusion, and the sensors with abnormal links are reported along with their verification records. When the number of abnormalities reaches a preset limit, the inspection conclusion of this round is marked as unreliable. In step S50, the inspection management platform processes the reported results as follows: the inspection conclusions of all reliable distribution boxes are archived, and their inspection intervals are increased when the verification pass meets the preset adjustment conditions; sensor maintenance work orders are generated for distribution boxes with abnormal links, and the inspection intervals are decreased; manual inspection work orders are generated for distribution boxes with unreliable inspection conclusions, and the acceptance of their remote alarms is suspended until the manual inspection is completed.

[0009] In some optional implementations, the excitation object can be at least one of a dehumidifier, a cooling fan, and internal lighting. The excitation action is to turn on and maintain it for 2 to 300 seconds before turning off. The verification table records the response range and response time limit for categories requiring change, the change range and upper limit of the change rate for categories requiring gradual change, and the allowable fluctuation range for categories requiring no change. The verification observation window consists of a baseline segment before switching and a response segment after switching; the comparison is based on the change in the response segment relative to the baseline segment. Inspection instructions are issued outside of peak power consumption periods in the transformer area. If an external event such as the door opening occurs within the observation window, the current round of verification is invalidated and re-arranged. Data from abnormal sensors within a traceability period prior to an anomaly determination is marked as pending verification. The increase in the inspection interval is conditional upon several consecutive rounds of full reliability and the adjustment does not exceed the preset range.

[0010] According to another aspect of the present invention, a remote inspection and management system for distribution boxes based on the Internet of Things is provided. The system includes an inspection management platform and an intelligent gateway installed in the distribution box. The inspection management platform includes a planning module and a handling module. The intelligent gateway includes an incentive execution module, a synchronous acquisition module, a verification and comparison module, and a conclusion generation module. Each module is responsible for the processing of generating and issuing inspection instructions, handling closed loop, incentive switching, synchronous acquisition through the observation window, verification table comparison, and conclusion generation and reporting.

[0011] According to another aspect of the present invention, an electronic device is provided, the electronic device including a processor and a memory, the memory storing a computer program, the computer program being executed by the processor to implement the IoT-based remote inspection and management method for distribution boxes as described in any of the above implementations.

[0012] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program for executing the IoT-based remote inspection and management method for distribution boxes described in any of the above implementations.

[0013] Based on the methods and systems provided in the above embodiments of the present invention, the acceptance of inspection data in each round is premised on a real physical stimulus verification. Stable false data such as constant values ​​and slow drift have nowhere to hide in the face of switching stimuli—if ​​the link is broken, it cannot be answered, thus blocking false detections from the mechanism. The stimulus source is taken from the existing load in the box, and the verification covers the entire link from the sensing element to the platform storage. Existing distribution boxes can obtain this capability by simply upgrading the gateway firmware, without the need to replace or add any sensing and verification hardware. The verification results simultaneously drive data governance and inspection rhythm—past data of abnormal channels are traced and marked to avoid bad data remaining in the ledger and misleading subsequent analysis. The inspection interval of healthy boxes is automatically opened, and maintenance resources are concentrated on boxes that really need attention. The faults of the stimulus object itself can also be identified and dispatched through cross-verification. The verification system will not miscalculate load faults to the sensor. Thus, the embodiments of the present invention transform remote inspection from "reading whatever is believed" to "verifying before believing," and the inspection conclusions have traceable and credible evidence for the first time.

[0014] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0015] Exemplary embodiments of the present invention can be more fully understood by referring to the following figures: Figure 1 A flowchart illustrating a remote inspection and management method for distribution boxes based on the Internet of Things according to an embodiment of the present invention; Figure 2This is a schematic diagram of the structure of a remote inspection and management system for distribution boxes based on the Internet of Things according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the data structure of the incentive response verification form; Figure 4 This is a schematic diagram illustrating the interaction process between the platform, gateway, stimulus object, and sensor during a round of inspection; Figure 5 This is a schematic diagram showing the comparison of outgoing current changes in two distribution boxes (one with a reliable link and one with an abnormal link) within the verification observation window in the embodiment. Figure 6 This is a schematic diagram showing the state of the inspection interval shifting between the standard setting, the upward setting, and the downward setting. Figure 7 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. The parameters such as excitation duration, response range, and inspection interval in the following implementations, as well as the current and temperature values ​​in the embodiments, are illustrative examples and do not constitute a limitation on the scope of protection of this invention. Those skilled in the art can adjust the corresponding parameters according to the actual configuration of the distribution box, and the adjusted implementation methods still fall within the scope of protection of this invention.

[0017] It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps described in these exemplary embodiments do not limit the scope of the invention. It should also be understood that, in the embodiments of the invention, "a plurality of" can refer to two or more, and "at least one" can refer to one, two, or more.

[0018] Those skilled in the art will understand that the terms "first," "second," etc., in the embodiments of the present invention are only used to distinguish different steps, devices, or modules, and do not represent any specific technical meaning, nor do they indicate a necessary logical order between them.

[0019] First, let's focus on explaining two fundamental terms that run throughout this article. The "existing controllable load" refers to electrical components already installed in the distribution box, controllable by the smart gateway, and serving the distribution box itself. Typical examples include dehumidifiers, cooling fans, and box lighting—these originally perform dehumidification, cooling, and lighting functions, but in this invention, they are used as a source of verification stimuli without adding any new components for verification purposes. The "stimuli response verification table" refers to a table pre-stored in the smart gateway, registering the expected response modes of each sensor according to the stimuli object. It answers the question, "What should each sensor in the box see after a certain load is connected?" It serves as the basis for verification and comparison, and its fields constitute a combination... Figure 3 Detailed explanation.

[0020] Exemplary methods Figure 1 This is a flowchart illustrating a remote inspection and management method for distribution boxes based on the Internet of Things (IoT) according to an embodiment of the present invention. Figure 1 As shown, the method includes steps S10 to S50, which are executed by the inspection management platform in cooperation with the smart gateway in the distribution box.

[0021] In step S10, the inspection management platform sends inspection instructions to the smart gateway via the IoT communication link according to the inspection plan. The inspection plan is automatically and continuously scheduled by the platform according to the current inspection interval of each distribution box, and the inspection times of adjacent boxes are staggered to avoid load overlap caused by multiple boxes in the same area being switched on and off simultaneously. The IoT communication link can be any form such as NB-IoT, Cat.1, or LoRa. Networking and transmission are mature technologies and will not be described in detail here. Unlike conventional remote inspection instructions, the inspection instructions of this method additionally carry verification incentive configuration. The verification incentive configuration refers to the description of the verification incentive used in this round of inspection, which at least specifies the incentive object and the duration of the incentive action. The incentive object refers to the existing controllable load that is switched on and off in this round, selected from the controllable load list registered in the distribution box—the platform maintains this list for each distribution box, registering the category, rated power, and rated current of each load for use when generating the verification incentive configuration. The duration is set from 2 to 300 seconds depending on the characteristics of the stimulus object: a shorter value, typically 5 seconds, is used for objects with immediate electrical response, such as interior lighting, that only test electrical channels; a longer value, typically 120 to 300 seconds, is used for objects that also test temperature channels, such as dehumidifiers, to allow sufficient time for heat dissipation. Optionally, the verification stimulus configuration can also specify the execution order of multiple stimulus objects, rotating different stimulus objects so that each sensor is covered in different rounds—for example, dehumidifiers are used to cover current and temperature channels in odd-numbered rounds, and interior lighting is used for rapid electrical verification in even-numbered rounds. The time taken for the lighting round is about 5% of that for the heater round, and the rotation arrangement takes into account both coverage and verification overhead. The inspection command is issued to avoid peak power consumption periods in the transformer area, for reasons explained in the comparison mechanism in step S30.

[0022] In step S20, the smart gateway switches the excitation object according to the verification excitation configuration and synchronously collects the output of each sensor within the verification observation window. The verification observation window refers to a continuous collection period covering the period before and after the switching. The baseline segment refers to the collection period within the observation window before the switching, used to obtain the background output of each sensor without excitation, typically 30 seconds; the response segment refers to the collection period after the switching, used to obtain the output under excitation, determined by adding a 10-second margin to the duration of the excitation action.

[0023] To address a common concern in engineering: will the increased switching for verification shorten the lifespan of the stimulating object? Take a dehumidifier heater as an example. Originally controlled by humidity linkage, it would switch on and off several times a day during the rainy season, sometimes even dozens of times. However, verification switching only occurs once per inspection interval. Based on a 7-day interval, this adds only about 52 switching events per year. Compared to the electrical lifespan of relays (tens of thousands of cycles) and the heater's own thermal inertia, this increase is negligible. The same applies to lighting and fans. The design of verification actions always adheres to the principle of "lightly touching" the components inside the enclosure. Synchronous data acquisition means that each sensor collects data at a uniform pace within the observation window: the current channel typically collects data once per second, the temperature channel once every 10 seconds, and the door sensor reports data instantly as an event. Each sample is uniformly stamped by the gateway, ensuring strict consistency between the baseline and response segments for each channel. The sampling frequency inside the observation window is an order of magnitude higher than normal monitoring. This is a unique encrypted acquisition during the verification period. Outside the window, the frequency drops back to normal. The additional acquisition overhead only occurs within a few minutes of each inspection interval, and its impact on the gateway's storage and power consumption is negligible. Each sensor includes at least an outgoing current sensor, an internal temperature sensor, and a door status sensor. If resources allow, humidity sensors, smoke sensors, etc., can also be included in the enclosure; this invention is not limited to these. Switching is executed by the relay output circuit of the smart gateway. This circuit is already an existing channel for the gateway to control the dehumidifier heater. This step simply changes the switching timing from humidity linkage to verification command triggering; there are no hardware modifications.

[0024] In step S30, the intelligent gateway compares the outputs of each sensor within the observation window with the excitation response verification table item by item. This is the core of the method's discrimination, combined with... Figure 3 Let's break down the verification form into its components. For example... Figure 3As shown, each item in the excitation response verification form is registered according to the excitation object. Fields include sensor identifier, expected response mode, judgment parameters, and response time limit. The expected response mode is divided into three categories: **"Change-oriented"**, meaning the sensor output should change in a definite direction and amplitude with the excitation; the response range refers to the acceptable range of output change required for this category; and the response time limit refers to the maximum time the change must occur from the start of the switch. For the excitation of connecting the dehumidifier heater, the output current sensor belongs to this category, and its response range is defined as 0.8 to 1.2 times the rated current of the dehumidifier heater, leaving margin for voltage fluctuations and measurement errors. **"Gradual-change"**, meaning the sensor output should show a change in a definite direction but at a gradual rate; the change range in its judgment parameters constrains the total change amplitude, and the upper limit of the change rate refers to the acceptable upper limit of the output change per unit time. Under the same excitation, the chamber temperature sensor belongs to the "gradual-change" category. The heat from the heater needs time to dissipate; the temperature should only rise slowly and not jump abruptly. A sudden jump indicates a problem with the channel. The "invariant" category refers to sensors whose output should not be affected by excitation; the permissible fluctuation range refers to the maximum allowable deviation of the invariant sensor's output from a reference. Door magnets belong to the invariant category, meaning that switching loads should not cause any change in the door's state.

[0025] The three modes together constitute a cross-verification: for the same stimulus action, some channels should respond, some should respond slowly, and some should remain silent. The three expectations are checked simultaneously, which verifies both "those that should respond will respond" and "those that should not respond will not respond randomly" - the latter can detect crosstalk between channels and switching faults such as misaligned data acquisition, which cannot be found by looking at a single channel.

[0026] The generation and maintenance of the stimulus response verification form do not require manual entry. The platform automatically generates the judgment parameters for each sensor based on the rated current and rated power registered in the controllable load list of each distribution box, using templates corresponding to the load categories. The response range for variable loads is obtained by multiplying the rated current by a fixed upper and lower coefficient. The response range for slowly variable loads is estimated based on the empirical relationship between load power and box volume, and then corrected by actual measurement in the first round of trial verification. The allowable fluctuation range for unchanging loads is taken from the sensor's factory accuracy specification. After the distribution box is replaced with a load or a sensor is added, the maintenance personnel only need to update the list entries, and the verification form will be regenerated and reissued. The maintenance cost is basically independent of the number of boxes. The specific execution of the comparison is based on the change of the response segment relative to the baseline segment: the average value of the baseline segment output is taken as the benchmark, and the difference between the output within the response segment and the benchmark is taken as the change. For variable loads, the change is checked to see if it falls within the response range and occurs within the response time limit. For slowly variable loads, the change and rate of change are checked. For unchanging loads, the output deviation from the benchmark is checked to see if it exceeds the allowable fluctuation range. The term "reliable link" refers to the sensor's entire path, from the sensing element through the acquisition circuit to the gateway for data entry, providing a true response that meets expectations under the current round of excitation. The term "abnormal link" refers to the output of the path not conforming to the expected response pattern, indicating a failure or degradation in the link. Sensors with matching outputs are considered to have reliable links, while those with mismatches are considered to have abnormal links. The original materials for each round of comparison—baseline values, response values, verification form entries, and judgment results—are stored as verification records at the gateway and carried with the report. Platform review and future dispute investigations are based on this record, making the verification itself auditable.

[0027] Here's a specific comparative calculation: In a certain round of verification, the average value of 30 samples of the baseline outgoing current was 12.46A. After connecting a 150W dehumidifier heater, the current in the response segment rose to 13.14A after 3 seconds, a change of 13.14-12.46=0.68A. The verification form recorded a response range of 0.55A to 0.82A (rounded to the nearest 0.8 to 1.2 times the rated current of 0.682A), with a response time limit of 5 seconds. 0.68A falls within this range, and the 3-second time limit is not exceeded, so the channel is considered reliable. After repeated comparisons and verifications by the inventor, the measured change in a normal link consistently falls within the middle of the range, while a failed link either remains unchanged or the change is completely inconsistent with the excitation magnitude. There is almost no ambiguity between the two types of results, and the comparison criteria do not require fine-tuning.

[0028] Another scenario prone to misjudgment needs to be rehearsed: the excitation period coincides with a sudden change in the actual load within the box—for example, the startup of a high-power device by a user in the observation window—the change in the outgoing current is the superposition of the excitation and load changes, which may exceed the response range and cause misjudgment. There are two countermeasures: First, the peak avoidance in step S10 has reduced such coincidences to a low probability; second, for the failure scenario corresponding to the change analogy, an automatic retry is arranged at intervals of several minutes. The actual link failure shows a stable performance during the retry, while the load coincidence will not hit the same observation window twice in a row. Only if the retry still does not match the result should an anomaly be judged.

[0029] This section explains two arrangements to ensure the effectiveness of comparisons. First, inspection instructions are issued outside of peak electricity usage periods: During peak periods, the load inside the distribution box fluctuates rapidly, and the background fluctuations in the outgoing current may overshadow the changes caused by the excitation. Avoiding peak periods ensures a stable baseline, making the excitation response significantly discernible. Typically, the issuance window is set between late night and early morning based on the distribution area's load curve. Second, if a box door is opened during the observation window, the current round of verification is invalidated and re-arranged: Opening the door indicates that someone is working on-site, and temperature and current may be disturbed. In this case, the comparison is meaningless, and invalidating and re-arranging is a conservative approach to prevent misjudgments.

[0030] In step S40, the smart gateway generates inspection conclusions based on the comparison results, categorized by scenario. When all sensors determine the link to be trustworthy, the inspection data collected in this round is accepted as a whole, and inspection conclusions are generated according to existing inspection discrimination rules—for example, whether the outgoing current exceeds the rated value, whether the box temperature exceeds the limit, and whether the box door is abnormally opened. These discriminations are mature aspects of remote inspection and will not be elaborated upon. The inspection conclusion refers to the conclusive record of the distribution box's operating status in this round of inspection, including at least the inspection time, the results of each discrimination item, and the overall status assessment, along with the pass status of this round of verification as evidence of the conclusion's credibility. The binding and archiving of conclusions and evidence is what distinguishes this method from traditional inspection logs that "only record conclusions without recording the basis." In the future, when tracing any historical conclusion, one can see what link status it was based on at that time. When some sensors determine that a link is abnormal, only the inspection data corresponding to sensors with reliable links are used to generate inspection conclusions. Abnormal sensors, along with their verification records (baseline values, response values, and violated expected patterns), are reported to the inspection management platform. Optionally, past data from abnormal sensors are marked with a tracing period. The preset tracing period refers to the time period from the time of the abnormality determination backwards, during which data needs to be re-examined. "Pending review" refers to the state where data, after being marked, is no longer used as reliable evidence for statistics and alarms, and is awaiting manual review or post-repair review. The preset tracing period does not exceed the interval between the sensor's most recent reliable link determination and the time of the abnormality determination—it was still good in the last verification; the problem could only have occurred between two verifications. Tracing back to the last reliable determination ensures that bad data is not overlooked and good data is not wrongly accused.

[0031] Another situation that appears similar to sensor anomalies but is actually different needs to be identified: a fault in the excitation object itself. If the dehumidifier heater burns out and there is no current when it is turned on, the output current channel will naturally show no change. Simply treating it as a discrepancy would misjudge a perfectly good sensor as a link malfunction. The identification criterion lies precisely in cross-verification—when the heater is truly faulty, both the current and temperature channels will simultaneously show "no response," and the probability of two independent sensor links failing simultaneously is far lower than a single point of failure in the excitation object. Therefore, when both the "expected change" and "expected gradual change" channels show discrepancies under the same excitation, the excitation object is prioritized as the faulty component. A repair work order for the excitation object is generated, and the system automatically re-verifies using alternative excitation objects from the list. If the re-verification matches, the reliability of each sensor link is maintained. This identification process allows the verification system itself to be verified. When the number of sensors identified as having link anomalies reaches a preset limit (typically half of the total number of sensors), the conclusion of this round of inspections is marked as unreliable. When most channels cannot answer, any remote conclusion loses its foundation. Honestly marking something as unreliable is more responsible than forcing a conclusion.

[0032] In step S50, the inspection management platform processes the reported results in three ways. For distribution boxes with all reliable links, the inspection conclusions are archived, and the inspection interval is increased when the verification meets the preset adjustment conditions. The preset adjustment conditions refer to the verification requirements that allow for a more relaxed inspection pace; typically, this means all sensors are deemed reliable for three consecutive rounds of inspections. The number of rounds can be adjusted from 2 to 6 based on the importance of the distribution box. The inspection interval is adjusted within a preset range, typically 3 to 14 days, with a standard range of 7 days—for continuously reliable distribution boxes, minimal disruption is required, and communication and platform resources are redirected elsewhere. For distribution boxes with link anomalies, a sensor maintenance work order is generated, and the inspection interval is lowered to the lower limit of the range. The maintenance work order includes the verification record of the abnormal sensor, allowing maintenance personnel to proceed directly to the specific channel upon arrival. For distribution boxes whose inspection conclusions are unreliable, a manual inspection work order is generated, and the remote alarms of that distribution box are temporarily suspended until the manual inspection is completed. Since the entire link is unreliable, its alarms may also be false alarms. Suspending the acceptance of alarms prevents maintenance personnel from being misled by false alarms. Suspending the acceptance of alarms does not mean discarding them: alarms of that box are still received and stored during the suspension period, but no action is taken. Once the manual inspection is completed and a subsequent round of verification confirms that all links are reliable, the platform automatically resumes acceptance of alarms. Alarms accumulated during the suspension period are identified and processed by maintenance personnel in conjunction with the results of the manual inspection. The restoration of alarm channels is also based on verification success, not on a manual "it should be fine." Optionally, the platform summarizes the verification success rate of each distribution box monthly. Boxes with consistently low success rates are suggested for overall renovation. The monthly summary also categorizes the failures by type, including terminal, probe, and gateway faults, providing a basis for spare parts reserves and renovation scheduling.

[0033] There is another robust arrangement at the communication level. NB-IoT links may be temporarily unavailable at certain times. When inspection commands fail to be issued, the platform handles the situation using a communication retry mechanism, which is standard practice. Even more noteworthy is the autonomous mode on the gateway side—the gateway stores the most recently issued verification incentive configuration and verification table locally. When the platform is unreachable at the locally scheduled inspection time, the gateway autonomously completes the switching, data collection, and comparison according to its local configuration. The conclusions and verification records are temporarily stored locally and retransmitted to the platform after the link is restored. The inspection cycle is not interrupted by communication fluctuations, and the platform processes the retransmitted data according to normal procedures.

[0034] Exemplary System Figure 2 This is a schematic diagram of the structure of a remote inspection and management system for distribution boxes based on the Internet of Things (IoT) according to an embodiment of the present invention. Figure 2 As shown, the system includes an inspection management platform 30 and an intelligent gateway 40 located in a power distribution box and connected to the inspection management platform 30 via an IoT communication link. The module division is illustrative of logical functions; in practice, they can be integrated into the platform software and gateway firmware respectively.

[0035] The inspection management platform 30 includes a planning module 31 and a handling module 32. The planning module 31 is used to support platform execution. Figure 1 Step S10 in the illustrated embodiment. The processing module 32 executes step S50. Specifically, the planning module 31 maintains the controllable load list, incentive response verification form template, and current inspection interval for each distribution box, and generates inspection instructions on a rolling basis according to the interval. The processing module 32 maintains the dispatch and closed-loop status of work orders and, in conjunction with the alarm subsystem, executes the suspension and recovery of remote alarms. The intelligent gateway 40 includes an incentive execution module 41, a synchronous acquisition module 42, a verification comparison module 43, and a conclusion generation module 44. The incentive execution module 41 executes the switching action in step S20. The synchronous acquisition module 42 executes the observation window acquisition in step S20. The verification comparison module 43 executes step S30. The conclusion generation module 44 executes step S40. The system provided in this embodiment is used to execute the method described in any of the aforementioned implementations, achieving the same technical effect. The processing mechanisms of each module will not be repeated.

[0036] Figure 4 This is a diagram illustrating the interaction process among various parties during a round of inspections. For example... Figure 4 As shown, the platform issues an inspection command carrying the verification incentive configuration. The gateway first collects the baseline, then switches the incentive object, and collects data synchronously in the observation window. Then, it compares the data with the verification table and generates an inspection conclusion. Finally, the conclusion and verification results are reported to the platform. The platform archives the data, adjusts the inspection interval, or dispatches work orders accordingly. One round of inspection ends in a closed loop.

[0037] Example The following combination Figure 5 and Figure 6 Taking the remote inspection and renovation of 38 outdoor low-voltage distribution boxes in an industrial park in southern China as an example, this paper reviews the entire execution process of this method. Each box is equipped with an outgoing current sensor, an internal temperature sensor, and a door magnet. Each box contains a 150W dehumidifier heater (rated current approximately 0.682A at 220V). The smart gateway is connected to the inspection management platform via NB-IoT. The verification incentive configuration uses the dehumidifier heater as the incentive object, with a duration of 120 seconds and a baseline period of 30 seconds. The incentive response verification form is as follows: Figure 3 The structure is as follows: the expected response mode for the outgoing current is variable, with a response range of 0.8 to 1.2 times the rated current (0.55A to 0.82A) and a response time of 5 seconds; the internal temperature is slow-change, with a change range of 0.2℃ to 2.5℃ within 120 seconds and a change rate upper limit of 0.6℃ / min; the door sensor is constant. The inspection interval is 3 to 14 days, with a standard interval of 7 days. An upward adjustment is required for three consecutive rounds of fully reliable inspections.

[0038] At 2:16 AM one night, distribution box number 17 underwent a routine inspection—this time coincided with the trough of the load curve for that area, with a stable baseline, making the excitation response easiest to discern. The average outgoing current in the baseline section was 12.46A, with a standard deviation of less than 0.03A over thirty samples. At 2:17 AM, the gateway was connected to the dehumidifier heater, and the current rose to 13.14A within 3 seconds. The change of 0.68A fell within the 0.55A to 0.82A range and was within a 5-second timeframe, indicating that the outgoing current path was reliable. Figure 5 As shown by the solid line, the step change is clean and crisp; within 120 seconds, the temperature inside the box slowly rises from 28.4℃ to 29.1℃, a change of 0.7℃ and a rate of change of approximately 0.14℃ / min, both within the parameters for slow-change scenarios, indicating the temperature channel link is reliable; the door magnet shows no state change throughout the process, and the output remains stable in the closed state, indicating the link is reliable. All three channels are reliable, and the data from this round of inspections is accepted as complete. The inspection conclusion is that the operation is normal, and supporting documentation of the successful verification of this round is archived. This box has been fully reliable for three consecutive rounds. The handling module increases its inspection interval from 7 days to 10 days, and after another 3 rounds, it will be increased to a maximum of 14 days. The entire process requires no human intervention.

[0039] That same night, the same verification was performed on distribution box number 23. Figure 5As shown by the dashed line, the average outgoing current of the baseline section of this box is 11.98A. After the dehumidifier heater is turned on, the current remains almost unchanged, and the average current during the response section is 11.99A, with a change of only 0.01A, far below the lower limit of the 0.55A range. The comparison between the two boxes on the same graph is obvious: the reliable link responds significantly to the excitation, while the failed link does not respond significantly to the excitation. The automatic retry results after 5 minutes remain unchanged, indicating that the outgoing current channel is abnormal; the temperature and door magnetic channels are verified to be consistent, indicating that the link is reliable. The conclusion generation module only accepts the temperature and door magnetic data to generate the inspection conclusion and reports the verification record of the current channel; the most recent reliable link determination for this channel was 14 days ago, so all current data within the previous 14 days are marked as pending verification. The platform dispatches a sensor repair work order to the maintenance team, with the work order directly stating "the outgoing current channel does not respond to an excitation of 0.68A". Maintenance personnel inspected the site and found a loose screw and oxidized contact surface on the terminal block connecting the secondary side of the current transformer to the gateway. After tightening, an on-site instant verification was triggered using a handheld terminal. The change in current channel excitation to the heater returned to 0.67A, falling within the acceptable range, and the excitation response returned to normal. The work order was closed based on this—whether the repair was truly fixed was determined by this excitation verification, not by manual judgment. Subsequent review of the data to be verified revealed that the current value returned from the loosening point remained around 11.98A—while the actual load of the box repeatedly exceeded 16A during this period. Although the 16A load was not exceeding the box's 25A rated capacity, relying on constant value data would have completely masked the subsequent overload. This failure, masked by a constant value, had no way of being exposed in traditional remote inspections. It was the excitation verification with a question in each round of this method that brought it to light on the 14th day. The outcome of the 14-day current data that was marked as pending review is also worth noting: After the maintenance was completed, the maintenance personnel cross-calculated the current data with the total electricity of the same box, and after confirming that there were no over-limit events during the period, the data was cancelled and released from the warehouse in batches. The current column in the platform ledger for this period was marked as "pending review and processed" rather than a smooth curve that could be mistaken for a real one - the closed loop of data governance was thus completed.

[0040] During the same period, another temperature channel issue was detected: In a round of verification of distribution box No. 31, the outgoing current channel responded normally to the heater excitation (change of 0.69A), but the temperature inside the box only rose from 26.8℃ to 26.9℃ in 120 seconds, a change of 0.1℃, which is below the lower limit of the 0.2℃ range. Based on the slow change category, the temperature channel link was judged to be abnormal. The automatic retry after a 5-minute interval yielded the same result, so it was classified as abnormal and a work order was dispatched. Upon arrival, it was found that the temperature probe had detached from its mounting bracket and fallen, hanging in the wiring trough at the bottom of the box—the probe itself was intact, but it was no longer measuring the temperature of the equipment area. This fault belongs to the type of "sensor not broken, location failure," which could not be detected by any built-in self-calibration (the probe self-test was completely normal). The excitation verification verifies the physical cause and effect that "you should feel heat when the heater is on," and the location failure was immediately apparent—this also clearly demonstrates that the verification object is the link and the entire installation, not just individual components.

[0041] As a comparative example 1, the verification process was eliminated, and the same record from distribution box No. 23 was directly replayed using traditional remote inspection data. Because the constant value returned by the current channel was statistically stable and showed no anomalies, all inspection records from the platform were normal for 14 consecutive days. The initial screening for model-related anomalies also failed to generate any alarms for this channel. The loose terminal might not be discovered until the next annual manual inspection or an accident. The two methods resulted in a difference of several months in the discovery time of the same failure. The difference lies not in the strength of the analysis algorithm, but in the presence or absence of a proactive, stimulating question-and-answer session.

[0042] As a comparative example 2, the same batch of distribution boxes were modified according to the self-calibration sensor route with built-in calibration sources: 38 boxes, totaling over a hundred sensors, needed to be completely replaced with models equipped with calibration circuits, requiring power outages for each box during the modification; moreover, the calibration signal was injected inside the sensor, and the loose terminal of number 23 and the detached probe of number 31 both occurred outside the sensor, yet the self-calibration process consistently concluded "sensor normal," and the two actual failures still slipped through the cracks. Because the calibration process did not cover the actual location of the failure, the hardware investment did not translate into corresponding detection capabilities—the difference in verification coverage is more decisive than the precision of the calibration method itself.

[0043] The overall performance of the 38 enclosures under this method for 6 months is as follows: approximately 1900 rounds of verification and inspection were performed, detecting 4 sensor link anomalies (2 cases of loose terminals, 1 case of probe detachment, and 1 case of gateway acquisition port failure). One excitation object failure was also identified (a dehumidifier heater burned out; after re-verification using alternative lighting excitation, the sensor link was confirmed to be intact). All cases were closed-loop managed with corresponding work orders. 31 enclosures maintained full reliability with a stable inspection interval of 14 days. The annualized number of on-site inspections decreased from 456 before the upgrade to 47, a reduction of approximately 90%. These 47 inspections were primarily for anomaly handling and spot checks. The workload of the upgrade itself is also noteworthy: all 38 enclosures were deployed remotely via firmware upgrades. The gateway upgrade and verification form distribution for each enclosure took approximately 10 minutes in total. No sensors were replaced and no wiring was altered throughout the process. The lightweight approach of upgrading existing equipment was fully validated in this batch of enclosures. The figures above are illustrative examples based on the park's operation log. Specific figures may vary depending on the size and condition of the container, but the incentive verification's ability to expose false inspections and the direction of saving inspection resources remain unchanged.

[0044] Figure 6 The diagram illustrates the migration relationship between the three inspection interval levels: the standard level moves to the upper level after three consecutive rounds of fully reliable monitoring; if a link anomaly occurs in any round, it falls to the lower level. After maintenance and closure and restoration of full reliability, the lower level returns to the standard level to re-accumulate data—the inspection rhythm is driven by verification results, requiring no manual intervention. The level distribution of 38 enclosures over 6 months also confirms the rationality of this rhythm: the 31 enclosures that remained at the upper level for a long time were the majority with excellent enclosure condition and link performance; the 5 enclosures that entered the lower level all corresponded to real maintenance events; none were unnecessarily tightened due to misjudgment, and the inspection density closely matches the actual needs of the enclosures.

[0045] In terms of deployment, this method only requires the smart gateway to have one relay output and firmware upgrade capability, which most in-service IoT gateways already possess. The incentive response verification form is generated by the platform according to the load nameplate parameters inside the box and then sent to the gateway. When adding a new box type, only its load parameters need to be registered. Modifications do not require changes to the sensors or the box wiring, making it practically scalable for the large number of existing distribution boxes. The inspection management platform can be deployed independently or embedded as a functional component of an existing power distribution maintenance master station. Verification results and inspection conclusions are available for access by the upper-level system through a standard interface. The link trust status can also be provided externally as a data quality label along with the inspection data, enabling downstream applications such as load analysis and status evaluation to use data with weighted reliability—the trustworthiness of sensor data now has a clear source. Exemplary electronic devices Figure 7 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Figure 7As shown, the electronic device 90 includes a processor 91 and a memory 92. The processor 91 can be a microcontroller in a smart gateway. The verification comparison only involves the mean, difference, and interval judgment, and the computational load is not a burden for any microcontroller. The memory 92 includes one or more combinations of volatile and non-volatile memory, storing a computer program. When executed by the processor 91, this computer program implements the IoT-based remote inspection and management method for distribution boxes described in any of the above implementations of this invention. The stimulus response verification table is also persistently stored in the memory 92. Optionally, the electronic device 90 also includes an input device 93 and an output device 94, such as a sensor interface, a relay output circuit, and an IoT communication module, with each component interconnected via a bus. For simplicity, Figure 7 Only some components relevant to this invention are shown.

[0046] Exemplary computer program products and computer-readable storage media In addition to the methods, systems, and electronic devices described above, embodiments of the present invention can also be computer program products, comprising computer program instructions. When executed by a processor, these computer program instructions cause the processor to perform the steps in the IoT-based remote inspection and management method for distribution boxes described in the preceding sections of this specification. The program code can be executed partly on the smart gateway and partly on the inspection and management platform side. Embodiments of the present invention can also be computer-readable storage media storing computer program instructions. When executed by a processor, these computer program instructions cause the processor to perform the steps in the aforementioned methods. The computer-readable storage medium can be random access memory, read-only memory, flash memory, or any suitable combination thereof. The program portion on the gateway side is written in firmware image form via a remote upgrade channel, and the program portion on the platform side is deployed as a service component. Both portions cooperate with the reporting interface according to the instructions described in this specification.

[0047] The above embodiments are merely one application of this method in the context of low-voltage distribution boxes. This method can be applied to any enclosed electrical equipment with controllable loads and various sensors built into the cabinet, such as outdoor ring network boxes and the low-voltage room of box-type substations. This invention is not limited to these.

[0048] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions described in the various embodiments of the present invention.

Claims

1. A remote inspection and management method for distribution boxes based on the Internet of Things, characterized in that, include: Step S10: The inspection management platform sends an inspection instruction to the smart gateway in the distribution box via the Internet of Things communication link according to the inspection plan. The inspection instruction carries the verification incentive configuration, which specifies the incentive object selected from the existing controllable load in the distribution box and the duration of the incentive action. Step S20: The smart gateway performs switching on the incentive object according to the verification incentive configuration, and synchronously collects the output of each sensor in the distribution box in the verification observation window before and after the coverage switching. Each sensor includes at least an outgoing current sensor, a box temperature sensor and a box door status sensor. Step S30: The smart gateway compares the output of each sensor in the verification observation window with the pre-stored excitation response verification table item by item. The excitation response verification table records the expected response mode of each sensor according to the excitation object. The expected response mode is divided into three categories: should change, should change slowly, and should remain unchanged. Sensors whose outputs match are considered reliable links, while those whose outputs do not match are considered faulty links. Step S40: When all sensors determine that the link is trustworthy, the inspection data collected in this round is accepted and an inspection conclusion is generated. When some sensors determine that the link is abnormal, only the inspection data corresponding to the reliable sensors in the link is used to generate the inspection conclusion, and the abnormal sensors along with the verification records are reported to the inspection management platform. When the number of sensors identified as having link anomalies reaches a preset limit, the conclusion of this round of inspections will be marked as unreliable. Step S50: The inspection management platform processes the reported results: archives the inspection conclusions of all reliable distribution boxes, and increases their inspection interval when the verification results meet the preset adjustment conditions. For distribution boxes with link anomalies, generate sensor maintenance work orders and reduce the inspection interval; For distribution boxes whose inspection conclusions are unreliable, a manual inspection work order is generated, and their remote alarms are suspended until the manual inspection is completed.

2. The method according to claim 1, characterized in that, The selectable range of the excitation object includes at least one of a dehumidifier, a cooling fan, and an internal lighting unit. The excitation action is to turn on the excitation object, maintain it for the specified duration, and then turn it off. The duration ranges from 2 seconds to 300 seconds.

3. The method according to claim 1, characterized in that, In the incentive response verification table, the expected response mode for the change category records the response amount range and response time limit, the expected response mode for the gradual change category records the change amount range and the upper limit of the change rate, and the expected response mode for the unchanged category records the allowable fluctuation range.

4. The method according to claim 3, characterized in that, For the excitation of turning on the dehumidifier heater, the expected response mode of the output current sensor is the change type, and its response range is defined as 0.8 to 1.2 times the rated current of the dehumidifier heater. The expected response mode of the temperature sensor inside the chamber is slow-change type; The expected response mode of the door status sensor is the invariant type.

5. The method according to claim 1, characterized in that, The verification observation window is composed of a baseline segment before switching and a response segment after switching. When comparing, the change in the response segment relative to the baseline segment is used as the sensor's output change. Inspection instructions should be issued at times that avoid peak electricity consumption periods in the area where the distribution box is located. If the box door is opened during the inspection observation window, the current round of inspections will be invalidated and re-arranged.

6. The method according to claim 1, characterized in that, Once a sensor is determined to have a link anomaly, the data of that sensor within a preset traceability period prior to the anomaly determination time is marked as pending review. The preset traceability period does not exceed the interval between the sensor's most recent link reliability determination and the anomaly determination time.

7. The method according to claim 1, characterized in that, The range of adjustment for inspection intervals is 3 to 14 days.

8. A remote inspection and management system for distribution boxes based on the Internet of Things (IoT), the system being used to implement the remote inspection and management method for distribution boxes based on the IoT as described in any one of claims 1 to 7, characterized in that, This includes an inspection management platform and a smart gateway installed in the distribution box and connected to the inspection management platform via an IoT communication link; The inspection management platform includes: a planning module, which generates and issues inspection instructions with verification incentive configurations according to the inspection plan. The verification incentive configurations specify the incentive objects selected from existing controllable loads in the distribution box and the duration of the incentive actions. The smart gateway includes: an incentive execution module, used to perform switching on the incentive object according to the verification incentive configuration; The synchronous acquisition module is used to synchronously acquire the outputs of each sensor in the power distribution box within the verification observation window before and after coverage switching; The verification and comparison module is used to compare the output of each sensor with the pre-stored excitation response verification table item by item to distinguish between reliable and abnormal links. The processing module is used to archive inspection conclusions based on the reports from the smart gateway, adjust inspection intervals, generate sensor maintenance work orders or manual inspection work orders, and suspend the acceptance of remote alarms from the corresponding distribution box when the inspection conclusions are unreliable. The conclusion generation module is used to accept inspection data based on comparison results, generate inspection conclusions, and report them to the inspection management platform.

9. An electronic device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, it implements the Internet of Things-based remote inspection and management method for distribution boxes as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the Internet of Things-based remote inspection and management method for distribution boxes as described in any one of claims 1 to 7.

Images