An internet-based remote monitoring system for electrical control cabinets

Abstract

The application belongs to the field of electrical automation, and relates to a remote monitoring system for an electrical control cabinet based on the Internet, which is composed of a cloud platform and an edge computing module. The cloud platform constructs a digital shadow model according to historical operation data of the control cabinet and issues the control logic pre-deployment to the edge computing module. The edge computing module collects electrical physical and environmental parameters in real time, calculates real-time safety thresholds and completes fault determination through local noise reduction and feature engineering processing combined with a dynamic compensation algorithm. When an abnormality is determined, the edge module directly drives an actuator through a local bus to implement power-off or isolation operation. During network interruption, time-stamped fault data is stored, and after network recovery, the data is synchronously transmitted to the cloud platform. The application effectively solves the defects of traditional remote monitoring systems, such as strong network dependence, large response time delay, high misoperation rate and difficult traceability, and meets the requirements of high reliability and intelligent monitoring in industrial sites.

Description

Technical Field

[0001] This invention belongs to the field of electrical automation, specifically relating to an Internet-based remote monitoring system for electrical control cabinets. Background Technology

[0002] Electrical control cabinets are key equipment in industrial power distribution and automation control; their safe and stable operation directly affects production continuity and equipment safety. Traditional local relay protection and manual inspection methods suffer from problems such as delayed response, high maintenance costs, and untimely fault handling. With the popularization of the Industrial Internet, existing remote monitoring systems mostly adopt a cloud-based centralized data acquisition, centralized judgment, and remote command issuance architecture, which generally has the following shortcomings:

[0003] It is highly dependent on network communication, and protection cannot be performed when the network is interrupted or delayed, which can easily lead to the expansion of the fault.

[0004] Cloud computing and transmission have high latency, and are insufficient for millisecond-level rapid fault response to phase-to-phase short circuits, overloads, etc.

[0005] Using a fixed protection threshold without dynamic correction based on operating conditions such as temperature and load rate can easily lead to false alarms or missed detections in high-temperature and heavy-load scenarios.

[0006] The lack of a self-protection mechanism for network outages and a high-precision time-stamped fault evidence chain makes it difficult to accurately trace the source and perform predictive maintenance.

[0007] CN117937768B discloses an intelligent electrical cabinet remote monitoring system, relating to the field of electrical cabinet monitoring technology. It includes an operation monitoring module for remotely collecting and preprocessing electrical cabinet operation data and equipment data using multimodal sensors, and simultaneously establishing a three-dimensional virtual model; a risk assessment module for evaluating the real-time operating status of the electrical cabinet based on operation data and taking corresponding measures; and a risk prediction module for predicting the future operating status of the electrical cabinet. This invention improves the accuracy of remote monitoring of electrical cabinets, significantly enhances system efficiency and security, and greatly strengthens the system's real-time performance and accident prevention capabilities by setting up multimodal sensors to remotely collect electrical cabinet operation data and geometric data, establishing a three-dimensional virtual model to evaluate and predict the operating status of the electrical cabinet, and formulating and implementing electrical cabinet operation data adjustment strategies based on the evaluation and prediction results. However, this existing technology still uses centralized cloud-based judgment, lacking edge-based local inference and autonomous protection during network outages; it does not deploy a digital shadow model and does not support adaptive thresholds based on historical trajectories; fault determination relies on fixed rules, resulting in a high rate of malfunction under high temperature and heavy load conditions; and it cannot perform power outages / isolation during network interruptions, making it impossible to completely retain and trace fault data.

[0008] CN220606314U discloses a remote data monitoring device for PLC equipment control cabinets, including a main body for remote data monitoring, a mounting bracket inside the main body, a placement through hole on one side of the main body and located on the side of the mounting bracket, a memory located inside the mounting bracket for reading and storing data from the PLC equipment control cabinet, and a locking component for disassembling and assembling the memory. This utility model, through a main body, a pad, a pull rod, a pull block, a spring, a baffle, and a push plate, uses a pull block to slide the pad within the mounting through hole, allowing the pad to move the locking block into the mounting through hole. Then, the push bracket pushes the push plate to move the memory. This solves the problem that existing equipment cannot easily repair or replace the storage device, nor can it easily disassemble and assemble the data storage device, leading to reduced stability of data transmission and affecting the stability and usability of data storage and retrieval. However, the existing technology relies entirely on the backend for all judgments; it lacks a digital shadow model and dynamic compensation algorithm, and the security threshold is fixed; it completely loses its protection capability when the network is disconnected; and it lacks fault transient data sealing and asynchronous synchronization mechanisms, which cannot meet the needs for rapid handling of millisecond-level faults.

[0009] In summary, existing technologies cannot simultaneously meet the requirements of millisecond-level response, autonomous protection against network outages, dynamic threshold self-adaptation, and full evidence chain tracing of faults, making it difficult to adapt to the stringent requirements of industrial sites for high reliability and intelligent monitoring of electrical control cabinets. Summary of the Invention

[0010] To address the shortcomings of the existing technologies, an Internet-based remote monitoring system for electrical control cabinets is provided to solve the aforementioned technical problems.

[0011] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is an Internet-based remote monitoring system for electrical control cabinets, including a cloud platform and an edge computing module;

[0012] The cloud platform constructs a digital shadow model based on the historical operating data of the electrical control cabinet, and distributes the digital shadow model to the storage unit of the edge computing module to realize the pre-deployment of control logic;

[0013] The edge computing module collects the electrical physical parameters and environmental parameters of the electrical control cabinet in real time, obtains the feature parameters through local noise reduction and feature engineering, and calculates the real-time safety threshold and performs fault determination through the local shadow model inference unit combined with the dynamic compensation algorithm.

[0014] When the judgment result is abnormal, the edge computing module directly drives the actuator through the local bus to perform power-off or isolation operations. After the network is restored, the decision records and fault waveforms with time stamps are synchronized to the cloud platform to achieve asynchronous state synchronization.

[0015] The aforementioned internet-based remote monitoring system for electrical control cabinets uses a cloud platform to construct a digital shadow model that fits the normal operating trajectory of the electrical control cabinet under different working conditions by mining historical current, voltage, frequency, and environmental indicators such as temperature and humidity inside the cabinet. This model is then packaged into a lightweight image file and sent to the edge computing module. The digital shadow model is either a support vector machine (SVM) model or a deep neural network (DNN) model, which is deployed after offline training and quantization compression in the cloud.

[0016] The aforementioned Internet-based remote monitoring system for electrical control cabinets uses an edge computing module that integrates a high-performance, reduced instruction set architecture multi-core processor as its edge gateway. The module periodically collects parameters at a fixed sampling interval of 10ms. Its data parsing unit uses the Kalman filter algorithm to smooth the original data stream, filter out electromagnetic interference, and extract high signal-to-noise ratio feature vectors.

[0017] The aforementioned internet-based remote monitoring system for electrical control cabinets includes a dynamic compensation algorithm that executes as follows: the edge computing module obtains the current temperature inside the cabinet. With load rate The local shadow model inference unit calculates correction coefficients based on temperature and load rate, and applies them to the preset safety threshold. Increase to obtain real-time security threshold The collected parameters are compared with the real-time safety threshold to determine whether the tripping logic is triggered.

[0018] The aforementioned internet-based remote monitoring system for electrical control cabinets specifies a real-time safety threshold that satisfies the following: ,in, , Calculate the correction coefficients for the dynamic compensation algorithm. For weighting factors in the digital shadow model.

[0019] The aforementioned Internet-based remote monitoring system for electrical control cabinets includes an edge computing module comprising a processor core, a storage unit, a data parsing unit, and a clock synchronization unit. The processor core runs a lightweight inference framework to load a digital shadow model. The clock synchronization unit aligns the time scale with the cloud platform via PTP or NTP protocols.

[0020] In the aforementioned Internet-based remote monitoring system for electrical control cabinets, when the network is interrupted, the edge computing module encapsulates the high-frequency sampled raw data from 100 milliseconds before the fault to 500 milliseconds after the fault into an evidence chain package and stores it in a non-volatile storage medium; after the network is restored, it is uploaded to the cloud platform through the breakpoint resume transmission protocol.

[0021] The aforementioned internet-based remote monitoring system for electrical control cabinets has the following evidence chain structure: The Header contains the device ID and global timestamp, and T is the high-frequency sampling sequence. : High-frequency sampling sequence before and after the fault, raw sampling data from 100ms before the fault to 500ms after the fault; Decisionlog: Local protection decision log, recording fault judgment and action information; CRC: Cyclic Redundancy Check code, used for data integrity verification;

[0022] The aforementioned Internet-based remote monitoring system for electrical control cabinets includes an actuator comprising an electromagnetic contactor, a shunt trip unit, and a disconnect switch; the edge computing module can disconnect the main circuit or isolate a damaged branch circuit within milliseconds via a local bus.

[0023] The aforementioned internet-based remote monitoring system for electrical control cabinets includes a local shadow model inference unit that monitors the three-phase current imbalance in real time. When the phase-to-phase short circuit or phase-out fault characteristic distribution is met, the system enters emergency response state.

[0024] The beneficial effects of the Internet-based remote monitoring system for electrical control cabinets in this invention are that it adopts edge local closed-loop control, skips network transmission and cloud protocol stack overhead, and reduces the fault handling time from the traditional >100ms to less than 10ms. It can effectively intercept rapidly evolving faults such as phase-to-phase short circuits and overloads, and prevent equipment damage and the spread of electrical fires.

[0025] By pre-deploying a lightweight digital shadow model in the cloud, the edge device can still complete the collection, judgment, and protection actions even when the network is interrupted and the system is in an "island state," thus solving the fatal flaw of traditional remote monitoring systems that fail when the network is paralyzed.

[0026] By combining operating conditions such as cabinet temperature and load rate, the safety threshold is adjusted in real time through a dynamic compensation algorithm, so that the threshold is close to the actual tolerance limit of the equipment, avoiding false alarms and unnecessary shutdowns in scenarios such as high temperature and heavy load.

[0027] Employing PTP / NTP high-precision clock synchronization, it automatically preserves the high-frequency sampling evidence chain from 100ms before to 500ms after the fault in the event of a failure. After the network is restored, it asynchronously synchronizes with the cloud, which can completely reproduce the fault transient and accurately locate the cause of the fault, providing data support for life prediction and preventive maintenance.

[0028] The cloud is responsible for big data modeling and offline training, while the edge is responsible for real-time inference and local protection. The off-grid computing power reduces the pressure on the cloud. The lightweight SVM / DNN model is quantized and compressed to adapt to edge hardware, and is compatible with multiple types of electrical control cabinets, making it applicable to a wider range of scenarios. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the overall architecture of the present invention. Detailed Implementation

[0030] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0031] like Figure 1 As shown, the internet-based remote monitoring system for electrical control cabinets includes a cloud platform and an edge computing module. The cloud platform constructs a digital shadow model of the electrical control cabinet using historical operating data and distributes this model to the storage unit of the edge computing module, enabling the pre-deployment of control logic. The edge computing module collects real-time electrophysical and environmental parameters within the electrical control cabinet, performs noise reduction and feature engineering locally to obtain feature parameters, and calculates real-time safety thresholds and performs fault determination using a local shadow model inference unit combined with a dynamic compensation algorithm.

[0032] The cloud platform undertakes big data analysis and high-level modeling tasks. Through in-depth analysis of historical electrophysical indicators such as current, voltage, and frequency, as well as environmental indicators such as temperature and humidity within the cabinet, a digital shadow model is constructed that can fit the normal operating trajectory of the electrical control cabinet under different working conditions. This model is packaged into a lightweight image file and distributed to the edge computing module via the internet. The edge computing module uses an edge gateway integrating a high-performance, reduced instruction set architecture (RISC) multi-core processor, communicating via a serial communication interface or an industrial Ethernet interface, according to T... S The raw values ​​of the current transformer and sensor are periodically acquired at a fixed sampling interval of 10ms. The data parsing unit in the edge computing module uses the Kalman filter algorithm to smooth the raw data stream in real time, filtering out complex electromagnetic interference from the field and extracting feature vectors with high signal-to-noise ratio. This application eliminates the negative impact of cloud communication latency on real-time monitoring by shifting the computational burden to the edge.

[0033] When the edge computing module determines that the result is abnormal, it directly drives the actuator through the local bus to perform power-off or isolation operations, and after the network is restored, it synchronizes the decision records and fault waveforms with time stamps to the cloud platform to realize the asynchronous state synchronization mechanism.

[0034] The edge computing module possesses complete autonomous decision-making power. When the local shadow model inference unit determines that the system is in a fault state or a warning state based on the current characteristic parameters, the digital output interface of the edge computing module instantaneously flips its level, directly triggering the actuator via the local bus. The actuator includes an electromagnetic contactor, a shunt trip unit, and a disconnect switch, thereby cutting off the main circuit or isolating the damaged branch within milliseconds. During network connection interruption, the edge computing module initiates an archiving mechanism, encapsulating the high-frequency sampled raw data from 100 milliseconds before to 500 milliseconds after the fault moment into an evidence chain packet and storing it in a non-volatile storage medium. Once the clock synchronization unit detects that the network handshake signal has recovered, the system automatically initiates the breakpoint resume protocol, uploading the archiving data packet to the cloud platform. This application, through the combination of local closed-loop control and asynchronous synchronization, ensures the system's survivability and data integrity under network paralysis conditions.

[0035] Thus, in the Internet-based remote monitoring system for electrical control cabinets, this application achieves a high degree of synchronization between the physical entity and the digital twin through the local deployment of the digital shadow model, providing an accurate benchmark for dynamic threshold determination; by utilizing the millisecond-level autonomous decision-making response at the edge, it avoids the non-deterministic latency of cloud interaction, greatly improving the success rate of intercepting rapidly evolving faults; and by using a dynamic compensation algorithm to adaptively correct the safety threshold under environmental conditions, it solves the problem of frequent malfunctions of traditional fixed thresholds under high temperature or overload conditions, and can meet the stringent requirements for high-reliability monitoring of control cabinets in industrial production.

[0036] In one embodiment, this application also provides an Internet-based remote monitoring system for electrical control cabinets, in which the process of the local shadow model inference unit executing the dynamic compensation algorithm includes:

[0037] The edge computing module obtains the current cabinet temperature through a sensor array. and current load rate ;

[0038] The local shadow model inference unit calculates the correction coefficient based on the cabinet temperature and load rate using a dynamic compensation algorithm. and the preset safety threshold The threshold is adjusted upwards to obtain the real-time safety threshold. ;

[0039] The edge computing module compares the real-time collected electrophysical parameters with the real-time safety threshold, and determines whether to trigger the tripping logic based on the comparison result.

[0040] In this embodiment, the dynamic compensation mechanism is the core of ensuring system intelligence. The digital shadow model embeds a weighted coefficient matrix for specific load characteristics. During the autonomous decision-making phase, if the edge gateway detects a heavy load condition with an internal cabinet temperature of 45 degrees Celsius and a load rate of 80%, the inference unit will invoke the dynamic compensation algorithm. The compensation logic follows the following physical mapping:

[0041] ;

[0042] in, These are the weighting factors in the digital shadow model. The resulting real-time security threshold... By dynamically adjusting the overload trip delay time or current threshold, false fault alarms caused by environmental heat accumulation are avoided. Simultaneously, the inference unit monitors the three-phase current imbalance in real time. This is used to identify the characteristic distribution of phase-to-phase short circuits and phase-loss faults, and the system enters an emergency response state when a threshold is reached. When the phase-to-phase short circuit or phase-out fault characteristic distribution as defined in the shadow model is met, the system enters an emergency response state.

[0043] In one embodiment, this application also provides an Internet-based remote monitoring system for electrical control cabinets, wherein the edge computing module in the system includes a processor core, a storage unit, a data parsing unit, and a clock synchronization unit.

[0044] The processor core is used to run a lightweight inference framework, which loads the digitized shadow model to perform a deterministic inference process;

[0045] The clock synchronization unit aligns its time stamp with the cloud platform using either Precision Time Protocol (PTP) or Network Time Protocol (NTP) to ensure that the locally recorded time stamp is accurate. With global standard time Consistency.

[0046] In this embodiment, the data parsing unit implements Kalman filtering at the FPGA or DSP hardware acceleration level, and the state transition equation of the filter is as follows:

[0047] ;

[0048] in, : Estimated characteristic value of current / voltage at time k The system state value at time k-1. For process noise, The state transition matrix describes the evolution of the system's states. : Control input matrix, The system control input at time k. The process noise is a system interference term that follows a Gaussian distribution. The processed net signal is input to the local shadow model inference unit. The digital shadow model preferably adopts a support vector machine (SVM) model or a deep neural network (DNN) model. These models are trained offline in the cloud and quantized, enabling them to run in the lightweight computing environment of the edge gateway. Since the model has fixed the operating trajectory of a specific control cabinet, the inference process does not rely on external real-time feedback and has extremely strong logical determinism.

[0049] In this embodiment, the operating status of the clock synchronization unit determines the accuracy of fault tracing. During system startup and operation, the clock synchronization unit periodically calibrates the local crystal oscillator frequency offset to ensure that each sampling point in the archived high-frequency sampling data packet has a precise nanosecond-level time stamp. When the network connection is restored, the edge computing module uploads the decision record with a unified time stamp, and the cloud platform performs subsequent lifetime prediction (RUL) and deep fault diagnosis analysis.

[0050] In one embodiment, the system provided in this application, after completing autonomous processing, seals the key evidence chain using a non-volatile storage medium. The structured representation of the evidence chain package is as follows:

[0051] ;

[0052] The Header contains the device ID and global timestamp, and T is the high-frequency sampling sequence. The data includes: a high-frequency sampling sequence before and after the fault, and raw sampling data from 100ms before the fault to 500ms after the fault; a decision log, a local protection decision log recording fault determination and action information; and a CRC (Cyclic Redundancy Check) code, used for data integrity verification. This sealing mechanism ensures that even after a system-wide power outage and restart, the physical process of the fault transient can still be completely reproduced. After receiving this data packet, the cloud platform compares the deviation between the standard trajectory in the digital shadow model and the measured fault trajectory to identify whether the specific fault mode is caused by cable aging, poor contact, or component breakdown.

[0053] Example 1:

[0054] Adaptive protection implementation plan for heavy-load and high-temperature conditions

[0055] This embodiment is applied to the electrical control cabinet for high-power motor drives in a continuous production workshop.

[0056] The edge computing module collects three-phase current, cabinet temperature and load rate in real time. When the cabinet temperature reaches 45℃ and the load rate reaches 80%, the local shadow model inference unit automatically starts the dynamic compensation algorithm.

[0057] The system calculates a correction coefficient based on temperature and load rate, and adaptively adjusts the preset current safety threshold upwards to obtain a real-time safety threshold. Under this condition, if the current fluctuates briefly, the system will not trigger a false trip; when the current exceeds the dynamically corrected threshold, the edge computing module drives the shunt trip unit to disconnect the main circuit within 8ms, completing the fault protection.

[0058] During network outages, the system fully saves high-frequency sampling data from 100ms before the fault to 500ms after the fault. After the network is restored, the data is automatically uploaded to the cloud. The cloud then combines the digital shadow model to complete fault type identification and lifespan prediction, achieving a combination of adaptive protection and predictive maintenance.

[0059] Example 2:

[0060] Example of isolated network operation and fault tracing

[0061] This embodiment is applied to the control cabinet of an outdoor box-type substation in an unstable network environment.

[0062] The edge computing module pre-stores lightweight digital shadow models distributed from the cloud. Even in an "island state" where the network is disconnected, it can still independently complete data acquisition, Kalman filtering noise reduction, local inference, and fault diagnosis.

[0063] When a phase-to-phase short-circuit fault occurs, the local shadow model inference unit identifies the fault characteristics within 3ms and directly drives the disconnecting switch to isolate the faulty branch, thus preventing the fault from escalating.

[0064] When a fault occurs, the system automatically encapsulates the fault waveform with nanosecond-level time stamps into an evidence chain packet and stores it in non-volatile FLASH memory. After the network is restored, the edge computing module synchronizes and calibrates the time using a PTP clock, and uploads the evidence chain packet to the cloud platform using breakpoint resume. The cloud platform compares the standard trajectory of the digital shadow model with the measured fault trajectory to quickly locate the phase-to-phase short circuit caused by aging of the line insulation, and generates maintenance suggestions to complete accurate fault tracing.

[0065] The Internet-based remote monitoring system for electrical control cabinets proposed in this application has the following significant advantages: it skips the encapsulation and transmission links of the network protocol stack, and reduces the local closed-loop processing time from more than 100ms in traditional remote monitoring to less than 10ms, effectively preventing the spread of electrical fires.

[0066] By using locally distributed digital shadow models, full-function monitoring was achieved in "isolated" environments, completely eliminating the dependence of remote control logic on WAN bandwidth and stability.

[0067] By combining dynamic compensation algorithms for environmental factors (temperature, load), the safety threshold is always aligned with the actual tolerance limit of the physical entity, significantly reducing the probability of non-faulty downtime.

[0068] By using high-precision time-stamp alignment and asynchronous data synchronization, a closed-loop evidence chain from edge acquisition to cloud analysis was constructed, providing a scientific basis for predictive maintenance of electrical assets.

[0069] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For system embodiments, since their logic is basically similar to that of the methods, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0070] The above provides a detailed description of an Internet-based remote monitoring system for electrical control cabinets provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. An internet-based remote monitoring system for electrical control cabinets, characterized in that, Includes cloud platform and edge computing modules; The cloud platform constructs a digital shadow model based on the historical operating data of the electrical control cabinet, and distributes the digital shadow model to the storage unit of the edge computing module to realize the pre-deployment of control logic; The edge computing module collects the electrical physical parameters and environmental parameters of the electrical control cabinet in real time, obtains the feature parameters through local noise reduction and feature engineering, and calculates the real-time safety threshold and performs fault determination through the local shadow model inference unit combined with the dynamic compensation algorithm. When the judgment result is abnormal, the edge computing module directly drives the actuator through the local bus to perform power-off or isolation operations. After the network is restored, the decision records and fault waveforms with time stamps are synchronized to the cloud platform to achieve asynchronous state synchronization.

2. The Internet-based remote monitoring system for electrical control cabinets according to claim 1, characterized in that, The cloud platform constructs a digital shadow model that fits the normal operating trajectory of the electrical control cabinet under different working conditions by mining historical current, voltage, frequency, and environmental indicators such as temperature and humidity inside the cabinet. The model is then packaged into a lightweight image file and sent to the edge computing module. The digital shadow model is a support vector machine (SVM) model or a deep neural network (DNN) model, which is deployed after offline training and quantization compression in the cloud.

3. The Internet-based remote monitoring system for electrical control cabinets according to claim 2, characterized in that, The edge computing module periodically collects parameters at a fixed sampling interval of 10ms. Its data parsing unit calls the Kalman filter algorithm to smooth the original data stream, filter out electromagnetic interference, and extract high signal-to-noise ratio feature vectors.

4. The Internet-based remote monitoring system for electrical control cabinets according to claim 1, characterized in that, The dynamic compensation algorithm is executed as follows: the edge computing module obtains the current temperature inside the cabinet. With load rate ; The local shadow model inference unit calculates correction coefficients based on temperature and load rate, and applies them to preset safety thresholds. Increase to obtain real-time security threshold The collected parameters are compared with the real-time safety threshold to determine whether the tripping logic is triggered.

5. The Internet-based remote monitoring system for electrical control cabinets according to claim 4, characterized in that, The real-time security threshold satisfies: ,in, , Calculate the correction coefficients for the dynamic compensation algorithm. For weighting factors in the digital shadow model.

6. The Internet-based remote monitoring system for electrical control cabinets according to claim 5, characterized in that, The edge computing module includes a processor core, a storage unit, a data parsing unit, and a clock synchronization unit; the processor core runs a lightweight inference framework to load a digital shadow model; the clock synchronization unit completes time synchronization with the cloud platform through PTP or NTP protocols.

7. The Internet-based remote monitoring system for electrical control cabinets according to claim 6, characterized in that, When the network is interrupted, the edge computing module encapsulates the high-frequency sampled raw data from 100 milliseconds before the fault to 500 milliseconds after the fault into an evidence chain package and stores it in a non-volatile storage medium. Once the network is restored, the data is uploaded to the cloud platform via the breakpoint resume protocol.

8. The Internet-based remote monitoring system for electrical control cabinets according to claim 5, characterized in that, The structure of the evidence chain is as follows: The Header contains the device ID and global timestamp, and T is the high-frequency sampling sequence. : High-frequency sampling sequence before and after the fault, raw sampling data from 100ms before the fault to 500ms after the fault; Decisionlog: Local protection decision log, recording fault judgment and action information; CRC: Cyclic Redundancy Check Code, used for data integrity verification.

9. The Internet-based remote monitoring system for electrical control cabinets according to claim 1, characterized in that, The actuator includes an electromagnetic contactor, a shunt trip unit, and a disconnect switch; the edge computing module can disconnect the main circuit or isolate the damaged branch circuit within milliseconds via the local bus.

10. The Internet-based remote monitoring system for electrical control cabinets according to claim 1, characterized in that, The local shadow model inference unit monitors the three-phase current imbalance in real time. When the phase-to-phase short circuit or phase-out fault characteristic distribution is met, the system enters emergency response state.

Images