Outdoor metering box remote operation control system based on Internet of Things

Abstract

The invention discloses an outdoor metering box remote operation control system based on the Internet of Things. The system comprises a multi-mode sensing device, an edge intelligent computing device and a cloud collaborative management platform. The multi-mode sensing device collects electrical parameters, environmental parameters and safety state parameters and sends the electrical parameters, the environmental parameters and the safety state parameters to the edge intelligent computing device through the controller local area network bus. And the edge intelligent computing device dynamically selects a communication link to upload data based on the link quality index, executes local anomaly detection through a compression neural network model deployed on an edge processor, and drives a circuit breaker to act when a preset control condition is met. And the cloud collaborative management platform is provided with an SM2 signature verification unit for performing SM2 algorithm signature verification on the remote control instruction, and the remote control instruction is forwarded to the edge intelligent computing device only when the verification is passed. According to the invention, high-reliability, intelligent, safe and controllable remote operation control of the outdoor metering box is realized.

Description

Technical Field

[0001] This invention belongs to the field of smart grid and Internet of Things integration technology, and relates to an IoT-based remote operation control system for outdoor metering boxes. Background Technology

[0002] With the deepening of smart grid construction, outdoor metering boxes, as key equipment at the end of the distribution network, undertake important functions such as electricity metering, load monitoring, and user-side management. Traditional outdoor metering boxes generally rely on manual inspections and lack real-time sensing capabilities for their operating status. Existing technologies, while some systems have introduced remote communication modules that can upload basic electrical parameters, offer only a single sensing dimension, covering only basic electrical parameters such as voltage and current, and failing to acquire environmental and security information such as temperature, humidity, water immersion, vibration, access control status, and anti-tampering status. This makes it difficult for the system to identify complex abnormal events such as electricity theft, equipment overheating, or unauthorized opening.

[0003] Furthermore, existing remote monitoring solutions mostly employ a single communication method, such as narrowband IoT or fourth-generation mobile communication. These are prone to communication interruptions under conditions of signal obstruction, electromagnetic interference, or extreme weather, resulting in data loss or control command failure. Even when some systems attempt dual-mode communication, their switching mechanisms are typically based on simple signal strength thresholds, failing to comprehensively consider bit error rate and latency, thus failing to guarantee the scientific validity and stability of link selection.

[0004] In terms of data processing, existing architectures mostly upload all raw data to the cloud, lacking intelligent analysis capabilities at the edge, resulting in high latency in anomaly response and consuming significant network bandwidth. Meanwhile, remote control commands generally lack strong authentication and integrity protection, posing a security risk of forgery or tampering. Although existing solutions employ common encryption protocols, they do not comply with national commercial cryptography standards, making it difficult to meet the compliance requirements of power monitoring systems.

[0005] Therefore, there is an urgent need for a remote operation control system for outdoor metering boxes that integrates multi-dimensional sensing, multi-link adaptive communication, edge intelligent decision-making, and national cryptographic security mechanisms to solve the technical problems of incomplete sensing, unreliable communication, untimely response, and unsafe control in existing technologies. Summary of the Invention

[0006] To address the problems existing in the background technology, this invention proposes a remote operation control system for outdoor metering boxes based on the Internet of Things.

[0007] The first aspect of this application provides an IoT-based remote operation control system for outdoor metering boxes, including: a multimodal sensing device, an edge intelligent computing device, and a cloud collaborative management platform;

[0008] The multimodal sensing device is used to collect electrical parameters, environmental parameters, and safety status parameters, and send the collected parameters to the edge intelligent computing device through the controller local area network bus.

[0009] The edge intelligent computing device is used to receive the parameters, dynamically select the communication link to upload data based on the link quality index, perform local anomaly detection through the compressed neural network model deployed on the edge processor, and drive the circuit breaker to perform opening or closing operations through the local control actuator when the preset control conditions are met.

[0010] The cloud collaborative management platform is used to receive uploaded data, construct a digital twin model of the metering box, and issue remote control commands verified by the SM2 algorithm signature.

[0011] Optionally, the multimodal sensing device includes an electrical parameter sensor, an environmental parameter sensor, and a safety status sensor; the electrical parameter sensor is used to collect three-phase voltage, three-phase current, active power, reactive power, and power factor; the environmental parameter sensor is used to collect temperature, humidity, water immersion status, and vibration acceleration; and the safety status sensor is used to collect access control switch status, anti-pry status, and electromagnetic shielding integrity status.

[0012] Optionally, the multi-link communication module in the edge intelligent computing device integrates a narrowband IoT communication chip, a fourth-generation mobile communication chip, a low-power wide-area network communication chip, and a power line carrier communication chip; the edge intelligent computing device is configured with a link quality assessment unit, which calculates link quality indicators based on the received signal strength, bit error rate, and end-to-end delay of each communication link, and selects the communication link with the highest link quality indicator for data upload.

[0013] Optionally, the formula used by the link quality assessment unit to calculate the link quality index is:

[0014] ;

[0015] in, Let i be the link quality index of the i-th communication link. Let be the received signal strength of the i-th communication link. Let be the bit error rate of the i-th communication link. Let be the end-to-end delay of the i-th communication link. , and Preset weighting coefficients and satisfying .

[0016] Optionally, the edge processor is deployed with a knowledge-distilled and compressed neural network model; the neural network model takes a multi-dimensional parameter sequence within a sliding time window as input and outputs the probability of an abnormal event occurring; when the probability is greater than a preset threshold, the edge processor triggers a local alarm and starts caching the original parameter data.

[0017] Optionally, the edge intelligent computing device stores a control rule library; the control rule library stores at least one control rule, and each control rule defines the mapping relationship between parameter conditions and circuit breaker actions; when the parameters collected in real time meet the parameter conditions in any control rule, the local control actuator generates the corresponding circuit breaker control signal.

[0018] Optionally, the cloud collaborative management platform is configured with an SM2 signature verification unit; the SM2 signature verification unit is used to perform SM2 algorithm signature verification on remote control commands; only when the verification is successful is the remote control command forwarded to the corresponding edge intelligent computing device.

[0019] Optionally, when the link quality index of the main communication link is lower than a preset threshold for two consecutive evaluation cycles, the edge intelligent computing device automatically switches to the backup communication link with the second highest link quality index and maintains the continuity of the communication session context.

[0020] Optionally, the multidimensional parameter sequence includes three-phase voltage, three-phase current, temperature, humidity, access control switch status, and vibration acceleration.

[0021] Optionally, when the edge intelligent computing device detects an abnormal event, it stores the original parameter data in a circular buffer and uploads the original parameter data first after the communication link is restored.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] This invention provides an IoT-based remote operation control system for outdoor metering boxes. Through multimodal sensing devices, it comprehensively collects electrical, environmental, and safety status parameters of the outdoor metering boxes, significantly improving the ability to identify complex anomalies. An edge intelligent computing device dynamically selects communication links based on link quality indicators, effectively avoiding the problem of easy interruption in complex environments with a single communication method, ensuring the continuity of data upload and command issuance. Through a compressed neural network model deployed on an edge processor, the system can complete anomaly detection locally in real time, without relying on cloud processing, greatly shortening response time. The local control actuator directly drives the circuit breaker when preset control conditions are met, achieving millisecond-level closed-loop protection and improving equipment and personnel safety. A circular buffer mechanism ensures that original parameter data is not lost when an anomaly occurs, and is prioritized for upload after communication is restored, providing complete evidence for post-event analysis. The cloud collaborative management platform uses an SM2 signature verification unit to strictly verify remote control commands, eliminating the risk of forged or tampered commands being executed, and meeting the safety and compliance requirements of the power system. Overall, this invention constructs a highly reliable, intelligent, and secure remote operation and control system for outdoor metering boxes, significantly improving the level of automated operation and maintenance and the ability to resist risks at the end of the power distribution network. Attached Figure Description

[0024] Figure 1 A schematic diagram of a remote operation control system for an outdoor metering box based on the Internet of Things (IoT) is shown in one embodiment of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] In one embodiment, such as Figure 1 As shown, an IoT-based remote operation control system for outdoor metering boxes is provided. This system includes a multimodal sensing device, an edge intelligent computing device, and a cloud collaborative management platform. Detailed descriptions of each functional module are as follows:

[0027] The multimodal sensing device is used to collect electrical parameters, environmental parameters, and safety status parameters, and send the collected parameters to the edge intelligent computing device through the controller local area network bus. The edge intelligent computing device is used to receive the parameters, dynamically select the communication link to upload data based on the link quality index, perform local anomaly detection through a compressed neural network model deployed on the edge processor, and drive the circuit breaker to perform opening or closing operations through the local control actuator when preset control conditions are met. The cloud collaborative management platform is used to receive the uploaded data, construct a digital twin model of the metering box, and issue remote control commands verified by the SM2 algorithm signature.

[0028] The multimodal sensing device is the core component of the system for achieving full-dimensional status sensing of the outdoor metering box. The device is integrally installed within the internal cavity and outer shell of the outdoor metering box and consists of three types of sensing components: electrical parameter sensors, environmental parameter sensors, and safety status sensors. These sensors are physically fixed and deployed in key locations such as near the electrical circuits inside the metering box, on the side walls, bottom, and door, ensuring the representativeness and real-time nature of the collected signals.

[0029] The electrical parameter sensor is directly connected to the main power supply circuit of the metering box to continuously collect three-phase voltage, three-phase current, active power, reactive power, and power factor. This sensor adopts an integrated design of a high-precision instrument transformer and analog-to-digital converter circuit, which can accurately reflect the electrical characteristics of the user's electrical load, providing a basic data source for subsequent identification of abnormal electricity consumption behavior.

[0030] Environmental parameter sensors include temperature sensors, humidity sensors, water immersion detectors, and vibration acceleration sensors. Temperature and humidity sensors are installed in the top space inside the metering chamber to monitor changes in the temperature and humidity of the chamber's microenvironment. The water immersion detector is located on the inner bottom of the chamber to detect rainwater or groundwater infiltration. The vibration acceleration sensor is fixed to the outer wall of the chamber or at the support connection point to sense external impacts or abnormal movement. These sensors together constitute a comprehensive monitoring capability for the metering chamber's operating environment.

[0031] The security status sensors include an access control switch, an anti-pry switch, and an electromagnetic shielding integrity detection coil. The access control switch uses a magnetic reed switch structure and is installed at the corresponding position on the door and inside the box to output a real-time signal indicating whether the door is open or closed. The anti-pry switch is a mechanical micro-trigger device embedded around the lock of the box, generating a trigger signal when the box is forcibly pried open. The electromagnetic shielding integrity detection coil surrounds the sensitive electronic area inside the metering box, and determines whether the shielding layer has been damaged by detecting changes in the coil circuit impedance, thereby identifying whether there is electromagnetic interference attack or illegal modification.

[0032] All the aforementioned sensors are connected to the edge intelligent computing device via a Controller Area Network (CLAN) bus. The CLAN bus employs differential signal transmission, providing strong resistance to electromagnetic interference and high communication reliability, making it suitable for complex outdoor electromagnetic environments. Each sensor samples synchronously according to a unified time base and sends the collected parameters to the edge intelligent computing device in a standardized data frame format. This data frame includes a sensor type identifier, parameter values, a timestamp, and a checksum, ensuring data integrity and traceability.

[0033] In one specific embodiment, when the outdoor metering box is deployed in a mountainous distribution area, the multimodal sensing device continuously collects signals of three-phase current imbalance, internal temperature rise, unauthorized door opening, and slight vibration. These parameters are transmitted in real time to the edge intelligent computing device via the controller area network bus, triggering a local anomaly analysis process, which in turn activates the circuit breaker trip protection and caches the original data. This embodiment demonstrates that the multimodal sensing device can effectively integrate electrical, environmental, and safety information, significantly improving the ability to identify complex faults or safety events, avoiding misjudgments based on single parameters, and thus enhancing the reliability and safety of the entire remote operation control system.

[0034] The edge intelligent computing device is the core hardware unit for realizing local intelligent decision-making and real-time control in the system of this invention. This device is installed inside an outdoor metering box, connected to a multimodal sensing device via a controller area network bus, and electrically coupled to the operating mechanism of the circuit breaker. It mainly consists of an edge processor, a local control actuator, a multi-link communication module, and a link quality assessment unit, all integrated on an industrial-grade embedded circuit board, possessing wide temperature range operating capability and high electromagnetic compatibility.

[0035] The edge intelligent computing device first receives various parameter data from the multimodal sensing device. These parameters include three-phase voltage, three-phase current, active power, reactive power, power factor, temperature, humidity, water immersion status, vibration acceleration, access control switch status, anti-tamper status, and electromagnetic shielding integrity status. All parameters are transmitted to the edge processor in the form of structured data frames via the controller area network bus, where they are parsed, verified, and time-aligned.

[0036] During the data upload phase, the edge intelligent computing device uses a link quality assessment unit to perform real-time evaluations of each communication link supported by the multi-link communication module. The multi-link communication module integrates narrowband IoT communication chips, fourth-generation mobile communication chips, low-power wide-area network communication chips, and power line carrier communication chips, each corresponding to different wireless or wired transmission channels. The link quality assessment unit continuously acquires the received signal strength, bit error rate, and end-to-end delay of each communication link and calculates the link quality index based on preset weighting coefficients. Subsequently, the device automatically selects the communication link with the highest link quality index for data upload, thereby ensuring the continuity and reliability of communication.

[0037] For local anomaly detection, the edge processor deploys a knowledge-distilled and compressed neural network model. This model is trained offline using a large amount of historical data and undergoes parameter pruning and quantization for resource-constrained environments, enabling it to run efficiently on the edge processor. The model takes a multi-dimensional parameter sequence within a sliding time window as input and outputs the probability of an anomaly occurring in the current time period. When this probability exceeds a preset threshold, the edge processor determines that an anomaly exists, triggers a local alarm mechanism, and simultaneously activates the raw parameter data caching function to temporarily store the relevant data in non-volatile memory.

[0038] At the control execution level, the edge intelligent computing device has a built-in control rule library, which stores multiple preset control rules. Each control rule clearly defines the mapping relationship between specific parameter conditions and circuit breaker actions. For example, when the three-phase current continuously exceeds 1.2 times the rated current for 5 seconds, or when the access control switch is open and the current time is not within the authorized maintenance period, the corresponding control conditions are met. Once the real-time collected parameters meet the conditions in any control rule, the local control actuator generates the corresponding control signal to drive the circuit breaker to perform opening or closing operations, realizing local closed-loop protection.

[0039] In one specific embodiment, an outdoor metering box was illegally opened, accompanied by a sudden increase in load. A multimodal sensing device detected that the access control switch had changed to open, and simultaneously, the three-phase current rose rapidly. Upon receiving this composite signal, the edge processor's neural network model outputs a high anomaly probability, while the control rule base matches the "illegal opening + overcurrent" combination. The local control actuator then issues a trip command, and the circuit breaker immediately trips, cutting off the power supply. Simultaneously, the device uploads the alarm information and cached raw data to the cloud collaborative management platform via a narrowband IoT link. This embodiment demonstrates that edge intelligent computing devices can quickly respond to complex security and operational events without cloud intervention, effectively preventing equipment damage or the escalation of electricity theft, and significantly improving the system's autonomous protection capabilities and response timeliness.

[0040] The multi-link communication module in the edge intelligent computing device is a key component for achieving highly reliable remote data transmission. This module integrates a narrowband IoT communication chip, a fourth-generation mobile communication chip, a low-power wide-area network (LPWAN) communication chip, and a power line carrier communication chip on the same circuit board. These four types of communication chips support different physical layer transmission mechanisms, covering various access methods such as cellular networks, unlicensed spectrum wireless networks, and power line wired communication. The narrowband IoT communication chip is suitable for wide-coverage, low-power remote transmission scenarios; the fourth-generation mobile communication chip provides high bandwidth and moderate latency communication capabilities; the LWAN communication chip operates in the Sub-1 GHz band, possessing strong penetration and long-distance transmission characteristics; and the power line carrier communication chip utilizes the power distribution lines connected to the metering box as the transmission medium, maintaining basic communication functions even when the wireless signal is completely lost.

[0041] To achieve intelligent selection of communication links, the edge intelligent computing device is equipped with a link quality assessment unit. This unit periodically reads low-level communication parameters from each communication chip, including received signal strength, bit error rate (BER), and end-to-end latency. Received signal strength, measured in decibels and milliwatts (dW), reflects the strength of the wireless link; BER indicates the proportion of erroneous bits during transmission, reflecting the link's stability; end-to-end latency, measured in milliseconds, refers to the time it takes for data to travel from the edge intelligent computing device to the cloud collaborative management platform for confirmation of reception, and is used to measure the link's real-time performance. The link quality assessment unit substitutes these three parameters into a preset link quality index calculation formula to obtain a comprehensive score for each communication link. This formula introduces three preset weighting coefficients, corresponding to the received signal strength, the reciprocal of the BER, and the reciprocal of the end-to-end latency, respectively, and their sum equals 1. Through this weighted calculation, the link quality assessment unit can objectively quantify the comprehensive performance of different links in the current environment.

[0042] After scoring each link, the link quality assessment unit compares the link quality index values ​​of all links and selects the communication link with the highest value as the current primary link for data upload. When the link quality index of the primary link remains below a preset threshold for more than two consecutive assessment cycles, the device automatically switches to the suboptimal link while maintaining uninterrupted communication session context, thus achieving seamless self-healing of the communication link.

[0043] In one specific embodiment, an outdoor metering box is deployed inside an underground power distribution well. Due to the obstruction of the concrete structure, narrowband IoT and 4G mobile communication signals are severely attenuated, with the received signal strength falling below -110 dBmW, resulting in a decline in link quality indicators. At this time, the low-power wide-area network communication chip, due to its lower operating frequency, maintains better penetration capability and has the highest link quality indicator, thus being selected as the primary link. Subsequently, a sudden rainstorm caused water accumulation in the well, completely interrupting wireless communication. However, the power line carrier communication chip can still transmit critical alarm data through the power distribution line. This embodiment demonstrates that the multi-link communication module, combined with the link quality assessment unit, can dynamically adapt to the optimal transmission path in complex and changing outdoor environments, effectively avoiding data loss or control failure caused by single-point communication failures, and significantly improving the communication robustness and availability of the entire remote operation control system.

[0044] The link quality assessment unit uses a specific mathematical expression to comprehensively reflect the multidimensional performance characteristics of the communication link when calculating link quality indicators. This expression is: .in, This represents the link quality index of the i-th communication link, used to characterize the overall communication capability of the link at present. This represents the received signal strength of the i-th communication link, measured in decibels and milliwatts, reflecting the power level of the wireless signal at the receiving end. The bit error rate of the i-th communication link is a dimensionless ratio that reflects the probability of errors occurring during data transmission. This represents the end-to-end latency of the i-th communication link, in milliseconds, which is the time it takes for data to be sent from the edge intelligent computing device to the cloud collaborative management platform for reception confirmation.

[0045] Since bit error rate and end-to-end delay are parameters that are better when they are smaller, while received signal strength is a parameter that is better when it is larger, therefore, in the formula... and Take the reciprocal so that its numerical properties are the same as... Consistency means that the larger the value, the better the link quality. , and These are preset weighting coefficients used to adjust the relative importance of the three types of parameters in the comprehensive evaluation. These three coefficients are set by the configuration file during system initialization, and their values ​​range from 0 to 1, and satisfy the following conditions: By adjusting the weighting coefficients, the performance requirements of different application scenarios can be adapted. For example, in scenarios emphasizing real-time control, the performance can be improved. The value of [value] is chosen to enhance the preference for low-latency links; in weak signal regions, it can improve [the performance]. The value is chosen to prioritize signal coverage.

[0046] In one specific embodiment, the edge intelligent computing device simultaneously enables narrowband IoT links and power line carrier links. Narrowband IoT links... It is -95 dBmW. It is 0.001. It takes 2000 milliseconds; power line carrier link has no... The parameters are different, but its equivalent signal strength is mapped to a negative 85 dB milliwatt equivalent through the signal-to-noise ratio. It is 0.01. The time is 100 milliseconds. The link quality assessment unit uses preset weights. =0.5、 =0.3、 =0.2, calculate the two links respectively. The calculation results show that although narrowband IoT links have a lower bit error rate, their high latency results in a lower overall score than power line carrier links. Therefore, the system selects power line carrier links to transmit time-sensitive data such as circuit breaker status changes. This embodiment illustrates that the formula can scientifically integrate signal strength, transmission reliability, and response speed, avoiding the one-sidedness of decision-making based on a single indicator, thereby achieving a more reasonable and robust communication link selection.

[0047] This link quality index calculation method enables edge intelligent computing devices to adapt to different environments. In complex outdoor deployment conditions, the performance of different communication technologies changes dynamically, and traditional fixed-priority switching mechanisms are prone to communication interruptions. This invention, however, effectively maintains the continuity of data uploads and the reachability of commands by quantitatively evaluating and dynamically optimizing links, significantly improving the overall communication reliability and operational stability of the remote operation control system.

[0048] The edge processor is the core computing unit in the edge intelligent computing device, on which a knowledge-distilled and compressed neural network model is deployed. This neural network model is learned from a large teacher model through knowledge distillation during the offline training phase, retaining its high ability to discriminate abnormal patterns while significantly reducing the number of parameters and computational complexity, enabling it to run efficiently in resource-constrained embedded environments. The model structure employs a lightweight combination of convolutional and fully connected layers, adapting to the memory capacity and computing power limitations of the edge processor.

[0049] The input to this neural network model is a sequence of multidimensional parameters within a sliding time window. The sliding time window has a fixed length and covers data from multiple consecutive sampling periods. The multidimensional parameter sequence includes various parameters from a multimodal sensing device, specifically including three-phase voltage, three-phase current, active power, reactive power, power factor, temperature, humidity, water immersion status, vibration acceleration, access control switch status, anti-tamper status, and electromagnetic shielding integrity status. These parameters are arranged in chronological order, forming a two-dimensional tensor, which serves as the model's input features. The model analyzes these time-series features through internal nonlinear transformations to identify implicit anomalous pattern correlations.

[0050] The output of the neural network model is a real number between 0 and 1, representing the probability of an abnormal event occurring in the current time period. This probability reflects the degree to which the system's operating state deviates from the normal pattern. A preset threshold is set during system initialization to determine whether an alarm should be triggered. When the output probability exceeds the preset threshold, the edge processor determines that a potential abnormal event exists and immediately performs two operations: first, it generates a local alarm signal, driving an audible and visual alarm or sending a high-priority alarm message to the cloud collaborative management platform; second, it activates the raw parameter data caching mechanism, writing the raw parameter data within the sliding time window and for several subsequent seconds into a circular buffer of non-volatile memory to ensure that critical evidence is not lost.

[0051] In one specific embodiment, an outdoor metering box experiences an abnormally high current during off-peak hours at night, simultaneously changing the access control switch from closed to open. A multimodal sensing device continuously collects this composite signal. The edge processor inputs a multidimensional parameter sequence from the past 60 seconds into a neural network model. The model identifies this combined feature as highly similar to historical electricity theft behavior, outputting an anomaly probability of 0.89. Since this value exceeds a preset threshold of 0.85, the edge processor immediately triggers a local alarm and completely caches the raw data, including the current waveform, the timing of the access control status change, and environmental parameters. After the communication link is restored, this cached data is preferentially uploaded to the cloud collaborative management platform for further analysis. This embodiment demonstrates that a knowledge-distilled and compressed neural network model can accurately identify complex anomalies at the edge, avoiding the latency caused by relying on cloud processing, while ensuring the integrity of post-event traceability through the raw data caching mechanism.

[0052] This technical solution enables the system to detect anomalies early and respond autonomously. Traditional threshold alarm methods struggle to identify hidden anomalies involving multiple coupled parameters. In contrast, this invention utilizes a lightweight artificial intelligence model to perform intelligent analysis on the terminal side, significantly improving the sensitivity and accuracy of detecting complex events such as electricity theft, equipment malfunction, or illegal intrusion. Simultaneously, it reduces reliance on network bandwidth, enhancing the intelligence level and operational security of the entire remote operation and control system.

[0053] The edge intelligent computing device internally stores a control rule base. This control rule base is stored in non-volatile memory in the form of structured data tables and can be configured or updated during system deployment or remote maintenance. The control rule base stores at least one control rule, and each control rule consists of a parameter condition section and a circuit breaker action section. The parameter condition section defines the logical combination relationship of one or more parameters, including parameter type, comparison operator, and threshold; the circuit breaker action section explicitly specifies the operation to be performed when the condition is met, including opening or closing.

[0054] The parameters involved in the conditions are derived from real-time data collected by the multimodal sensing device, including three-phase current, three-phase voltage, temperature, access control switch status, and anti-tamper status. Comparison operators include logical judgment methods such as greater than, less than, equal to, not equal to, and duration greater than. For example, the parameter condition for one control rule can be defined as "the three-phase current is greater than 1.2 times the rated current and the duration of this state is greater than 5 seconds"; another control rule's parameter condition can be defined as "the access control switch status is equal to open and the current system time is not within the preset maintenance time period". Circuit breaker actions correspond to "perform opening operation" or "prohibit closing operation".

[0055] When the edge intelligent computing device is running, the local control actuator periodically obtains the latest parameters from the multimodal sensing device and matches them against the parameter conditions in the control rule base one by one. The matching process uses a real-time logic judgment engine to evaluate the conditional expression of each rule. Once the parameter condition of any control rule is met, the local control actuator immediately generates the corresponding circuit breaker control signal. This signal is amplified by an opto-isolation circuit and drives the electromagnetic operating mechanism of the circuit breaker to complete the opening or closing action. The entire process is completed locally in a closed loop, without waiting for cloud commands, ensuring timely and reliable response.

[0056] In one specific embodiment, a short-circuit fault occurred in the user line connected to an outdoor metering box, causing the three-phase current to instantaneously rise to 1.5 times the rated value. The edge intelligent computing device detected that this current value exceeded the overcurrent threshold set in the control rules, and the duration exceeded 5 seconds. The corresponding rule in the control rule library was triggered, and the local control actuator immediately output a high-level trip signal, causing the circuit breaker to disconnect the faulty circuit within hundreds of milliseconds. At the same time, another rule detected that the box door was opened during an unauthorized period, which also independently triggered the trip protection. The two abnormal events were handled by different rules and did not interfere with each other. This embodiment shows that the control rule library supports parallel judgment of multiple conditions and scenarios, and can implement accurate and rapid local protection for different safety or operational risks.

[0057] This technical solution enables rule-based autonomous decision-making. Compared to traditional control systems that rely solely on remote commands, this invention effectively addresses emergency situations such as communication interruptions or cloud latency by locally storing and executing control rules. The configurable nature of the control rule base also supports the deployment of differentiated protection strategies based on different regions and user types, enhancing the system's adaptability and security. Overall, this mechanism strengthens the proactive protection capabilities of outdoor metering boxes in unattended environments, ensuring the safety of end-point equipment and personnel in the power distribution network.

[0058] The edge intelligent computing device continuously monitors the link quality indicators of the main communication link during operation. These indicators are calculated by the link quality assessment unit according to a preset cycle, each cycle being called an assessment cycle. The preset threshold is a fixed value set during system initialization and is used to determine whether the current communication link meets reliable transmission requirements. When the link quality indicator of the main communication link falls below this preset threshold, and this state occurs consecutively in two adjacent assessment cycles, the edge intelligent computing device determines that the main communication link has entered an unstable or failed state.

[0059] At this point, the edge intelligent computing device initiates a link switching process. This process first recalculates the link quality indicators (MMI) of all available communication links, excludes the current primary link, and selects the link with the highest MMI as the backup communication link. Subsequently, the device seamlessly migrates the data upload task to this backup communication link. During the switching process, the edge intelligent computing device maintains the state information of the upper-layer communication protocol, including session identifier, sequence number, acknowledgment status, and incomplete data fragments, ensuring the continuity of the communication session context. The cloud collaborative management platform does not need to be aware of the underlying link change at the receiving end and can continue to process subsequent data packets according to the original logic, avoiding connection interruptions or data retransmissions due to link switching.

[0060] In one specific embodiment, an outdoor metering box is deployed in a power distribution room in an urban tunnel, with an initial main communication link being a fourth-generation mobile communication link. Due to signal fluctuations within the tunnel, the link quality index of this link drops below a preset threshold in the first evaluation cycle. The edge intelligent computing device does not switch over and continues to observe. In the second consecutive evaluation cycle, the link quality index still has not recovered and remains below the preset threshold. The device then activates the low-power wide-area network communication chip, using it as a backup communication link. During the switchover process, the circuit breaker status change message that is being uploaded continues to be sent through the new link, with its sequence number continuous with the previous message. The cloud collaborative management platform receives and correctly parses the message. The entire process is imperceptible to the user, and the control command channel remains unobstructed.

[0061] This mechanism effectively solves the problem of single communication links being easily interrupted in complex electromagnetic or geographical environments. By introducing a dual-cycle acknowledgment mechanism, it avoids erroneous handovers caused by momentary interference; and by maintaining the continuity of the communication session context, it ensures the reliable transmission of critical control and alarm data. Overall, this design significantly improves the communication resilience and service availability of remote operation and control systems in weak or dynamic network environments.

[0062] When an abnormal event is detected, the edge intelligent computing device activates a data caching mechanism to ensure the integrity of critical operational information. Abnormal event detection is performed by a knowledge-distilled and compressed neural network model deployed on the edge processor, or it can be triggered by logical judgments in the control rule base. Once an abnormal event is determined to have occurred, the edge intelligent computing device immediately writes the raw parameter data collected by the multimodal sensing device into a preset circular buffer in non-volatile memory. This circular buffer has a fixed storage capacity and uses a first-in, first-out (FIFO) strategy to manage data. When the buffer is full, new data automatically overwrites the oldest data, ensuring that the complete original records from the most recent period are always retained.

[0063] The raw parameter data includes three-phase voltage, three-phase current, active power, reactive power, power factor, temperature, humidity, water immersion status, vibration acceleration, access control switch status, anti-tamper status, and electromagnetic shielding integrity status. This data is organized by sampling period using timestamp alignment, preserving complete temporal characteristics and providing high-fidelity data for subsequent cloud-based analysis. During caching, the edge intelligent computing device simultaneously marks this data segment as high-priority event data and records the anomaly type, trigger time, and associated control action information.

[0064] When a communication link is interrupted or degraded, caching operations continue independently of the upload process. Once the multi-link communication module detects that the link quality indicators of any communication link have recovered to an usable level, the edge intelligent computing device prioritizes the upload of high-priority event data marked in the circular buffer. The upload order follows a dual principle of chronological order and event severity, ensuring that the earliest abnormal data is delivered to the cloud collaborative management platform first. Upload tasks for ordinary operational data are temporarily downgraded or suspended during this period to ensure that bandwidth resources are allocated to critical events.

[0065] In one specific embodiment, an outdoor metering box was struck by lightning, causing a momentary communication interruption, along with a sudden rise in internal temperature and distortion of the current waveform. The edge processor, using a neural network model, identified the combined anomaly and immediately activated its cache, storing raw parameter data from the 10 minutes before and after the lightning strike into a circular buffer. For the next 30 minutes, all wireless links were unavailable. Once the weather improved, the low-power wide-area network (LPWAN) communication link signal was restored, and link quality indicators returned to normal. The edge intelligent computing device then prioritized uploading the cached anomaly event data packet, along with an event tag. Upon receiving this data, the cloud collaborative management platform, using a digital twin model, reconstructed the fault process, assisting maintenance personnel in accurately locating the equipment damage points.

[0066] This mechanism effectively solves the problem of critical data loss during communication interruptions. Local persistent storage is achieved through a circular buffer, combined with a priority scheduling strategy to ensure no abnormal evidence is missed. Even in extreme scenarios with prolonged network connectivity loss, the system retains complete event backtracking capabilities. Overall, this design significantly enhances the data reliability and post-event analysis capabilities of the remote operation and control system, providing solid support for the safe operation of the power grid.

[0067] The cloud-based collaborative management platform is a core component of the system of this invention, enabling centralized management and secure command distribution. Deployed in the power company's private data center or controlled cloud environment, the platform is internally configured with an SM2 signature verification unit. This SM2 signature verification unit is a dedicated security processing module built based on the SM2 elliptic curve public key cryptography algorithm released by the State Cryptography Administration, used to verify the integrity and authenticity of all remote control commands initiated by maintenance terminals or management interfaces.

[0068] When a remote control command is generated, the authorized operator digitally signs the command content using the corresponding SM2 private key. The signing process includes hashing the command data and encrypting the hash value using the SM2 private key to generate a unique digital signature. This signature, along with the original command, is packaged into a secure command packet and sent to the cloud collaborative management platform. Upon receiving the secure command packet, the SM2 signature verification unit first extracts the digital signature and the plaintext command. Then, using the SM2 public key of the corresponding edge intelligent computing device, which is pre-registered and stored in the platform's security database, it executes the verification process defined by the SM2 algorithm. The verification includes signature format compliance, hash value consistency, and public key validity.

[0069] A remote control command is deemed legitimate and trustworthy only when the SM2 signature verification unit confirms that the digital signature is valid and matches the command content. At this point, the cloud collaborative management platform forwards the command to the target edge intelligent computing device through the established secure communication channel. If verification fails, including in cases of invalid signature, public key mismatch, or command tampering, the SM2 signature verification unit immediately discards the command, records a security audit log, and simultaneously returns a verification failure message to the operator to prevent the execution of illegal or forged commands.

[0070] In one specific embodiment, maintenance personnel issue a remote closing command to restore power to a user via the dispatch master station. This command is signed on the master station side using the SM2 private key bound to the metering box's device certificate. After the command is transmitted to the cloud collaborative management platform, the SM2 signature verification unit retrieves the SM2 public key certificate corresponding to the metering box and executes the standard SM2 verification algorithm. Upon successful verification, the platform sends the closing command to the edge intelligent computing device of the metering box via a DTLS encrypted channel. The device receives the command and performs local signature verification again before finally executing the closing operation. Throughout this process, even if an attacker intercepts and attempts to replay or tamper with the command, it will be blocked because it cannot pass the SM2 signature verification.

[0071] This mechanism fundamentally eliminates the risk of remote control commands being forged, replayed, or tampered with by a man-in-the-middle. By adopting the national commercial cryptography standard SM2 algorithm, it ensures that the system complies with relevant regulations for the security protection of power monitoring systems. The SM2 signature verification unit serves as the sole secure gateway for command access, enabling the entire remote operation and control system to possess end-to-end identity authentication and data integrity assurance capabilities, significantly improving the security level and anti-attack capabilities of the distribution network's end-point control.

[0072] The modules in the aforementioned IoT-based remote operation control system for outdoor metering boxes can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, allowing the processor to call and execute the corresponding operations of each module.

[0073] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A remote operation control system for outdoor metering boxes based on the Internet of Things, characterized in that, include: Multimodal sensing devices, edge intelligent computing devices, and cloud collaborative management platforms; The multimodal sensing device is used to collect electrical parameters, environmental parameters, and safety status parameters, and send the collected parameters to the edge intelligent computing device through the controller local area network bus. The edge intelligent computing device is used to receive the parameters, dynamically select the communication link to upload data based on the link quality index, perform local anomaly detection through the compressed neural network model deployed on the edge processor, and drive the circuit breaker to perform opening or closing operations through the local control actuator when the preset control conditions are met. The cloud collaborative management platform is used to receive uploaded data, construct a digital twin model of the metering box, and issue remote control commands verified by the SM2 algorithm signature.

2. The remote operation control system for outdoor metering boxes based on the Internet of Things as described in claim 1, characterized in that, The multimodal sensing device includes an electrical parameter sensor, an environmental parameter sensor, and a safety status sensor; the electrical parameter sensor is used to collect three-phase voltage, three-phase current, active power, reactive power, and power factor; the environmental parameter sensor is used to collect temperature, humidity, water immersion status, and vibration acceleration; the safety status sensor is used to collect access control switch status, anti-pry status, and electromagnetic shielding integrity status.

3. The remote operation control system for outdoor metering boxes based on the Internet of Things according to claim 1, characterized in that, The multi-link communication module in the edge intelligent computing device integrates a narrowband IoT communication chip, a fourth-generation mobile communication chip, a low-power wide area network communication chip, and a power line carrier communication chip. The edge intelligent computing device is equipped with a link quality assessment unit, which calculates link quality indicators based on the received signal strength, bit error rate, and end-to-end delay of each communication link, and selects the communication link with the highest link quality indicator for data upload.

4. The IoT-based remote operation control system for outdoor metering boxes according to claim 3, characterized in that, The formula used by the link quality assessment unit to calculate the link quality indicators is: ； in, Let i be the link quality index of the i-th communication link. Let be the received signal strength of the i-th communication link. Let be the bit error rate of the i-th communication link. Let be the end-to-end delay of the i-th communication link. , and Preset weighting coefficients and satisfying .

5. The IoT-based remote operation control system for outdoor metering boxes according to claim 1, characterized in that, The edge processor is equipped with a knowledge-distilled and compressed neural network model; the neural network model takes a multi-dimensional parameter sequence within a sliding time window as input and outputs the probability of an abnormal event occurring; when the probability is greater than a preset threshold, the edge processor triggers a local alarm and starts caching the original parameter data.

6. The remote operation control system for outdoor metering boxes based on the Internet of Things according to claim 1, characterized in that, The edge intelligent computing device stores a control rule library; the control rule library stores at least one control rule, and each control rule defines the mapping relationship between parameter conditions and circuit breaker actions; when the parameters collected in real time meet the parameter conditions in any control rule, the local control actuator generates the corresponding circuit breaker control signal.

7. The IoT-based remote operation control system for outdoor metering boxes according to claim 1, characterized in that, The cloud collaborative management platform is configured with an SM2 signature verification unit; the SM2 signature verification unit is used to verify the remote control command using the SM2 algorithm signature; only when the verification is successful is the remote control command forwarded to the corresponding edge intelligent computing device.

8. The remote operation control system for outdoor metering boxes based on the Internet of Things according to claim 1, characterized in that, When the link quality index of the main communication link is lower than a preset threshold for two consecutive evaluation cycles, the edge intelligent computing device automatically switches to the backup communication link with the second highest link quality index and maintains the continuity of the communication session context.

9. The IoT-based remote operation control system for outdoor metering boxes according to claim 5, characterized in that, The multidimensional parameter sequence includes three-phase voltage, three-phase current, temperature, humidity, access control switch status, and vibration acceleration.

10. The IoT-based remote operation control system for outdoor metering boxes according to claim 1, characterized in that, When the edge intelligent computing device detects an abnormal event, it stores the original parameter data in a circular buffer and uploads the original parameter data first after the communication link is restored.

Images