A remote control system of integrated distribution box based on 5G communication

Abstract

The application relates to the technical field of data processing, in particular to a comprehensive power distribution box remote control system based on 5G communication, which comprises a data processing module, an analysis module, a remote control module, a control judgment module, a trimming adjustment module and a constraint adjustment module; the data processing module collects operation data of the comprehensive power distribution box, pre-processes the operation data to obtain optimized operation data, and transmits the optimized operation data to a cloud platform through a 5G network; the analysis module obtains operation characteristics by performing feature extraction on the optimized operation data, constructs a machine learning model according to the operation characteristics, and obtains analysis results by analyzing the operation data; the remote control module generates a control scheme according to the analysis results; the control judgment module determines the effectiveness of the remote control of the comprehensive power distribution box according to control characteristic values; the trimming adjustment module determines an invalid characteristic pruning threshold of the machine learning model according to a redundancy rate of the operation characteristics; and the constraint adjustment module determines a time sequence neighborhood coupling constraint coefficient of the operation characteristics according to a transmission time delay jitter amplitude of the optimized operation data. The application improves the analysis accuracy of the operation data of the comprehensive power distribution box.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a remote control system for integrated power distribution boxes based on 5G communication. Background Technology

[0002] In existing technologies, 5G communication has been widely applied in remote management and control scenarios for power distribution equipment. Leveraging its high-speed transmission advantage, it enables remote acquisition and network monitoring of integrated power distribution box operation data. Currently, conventional 5G network systems for power distribution boxes only possess basic data upload, simple statistics, and threshold alarm functions, lacking refined time-series analysis and data quality verification mechanisms. Furthermore, 5G multi-path transmission is prone to latency jitter and out-of-order data packets. Existing solutions do not address time-series disorder issues with realignment compensation, redundant feature pruning, or time-series constraint control strategies. Out-of-order data directly participates in computation and analysis, easily leading to feature dimension redundancy and data parsing distortion. Moreover, traditional power distribution management and control systems lack a quantitative judgment system for transmission effectiveness, feature extraction effectiveness, and remote control effectiveness. They cannot perform layered tracing of issues such as data distortion, network disturbances, and inefficient model inference. Various operating parameters are mostly statically configured, making it difficult to adapt to complex power grid conditions and dynamically fluctuating communication environments, resulting in insufficient accuracy in the analysis of integrated power distribution box operation data.

[0003] Chinese Patent Publication No. CN113972743A discloses a comprehensive power distribution box and its network system based on 5G technology. The system includes a box body and power distribution devices and a network control device installed within the box body. The network control device includes a controller, a data collection module, a data processing module, a data analysis module, and a data upload module. The controller is equipped with a 5G communication module connected to a 5G base station. The data collection module includes an internal detection module and an external data receiving module. The internal detection module includes an input voltage and current detection module, a multi-channel output voltage and current detection module, and a box environment detection module. The input voltage and current detection module detects the voltage and current of the input lines, and the multi-channel output voltage and current detection module detects the voltage and current of multiple output lines. Therefore, the comprehensive power distribution box and its network system based on 5G technology suffer from insufficient accuracy in analyzing operational data because the data analysis only includes unit-time and overall analysis, lacking feature extraction and intelligent predictive analysis based on historical operating patterns. Summary of the Invention

[0004] To address this issue, the present invention provides a remote control system for integrated distribution boxes based on 5G communication, which overcomes the problem in the prior art where data analysis only includes unit time and overall analysis, lacking feature extraction and intelligent predictive analysis of operating data in combination with historical operating patterns, resulting in insufficient accuracy in the analysis of integrated distribution box operating data.

[0005] To achieve the above objectives, the present invention provides a remote control system for integrated distribution boxes based on 5G communication, comprising: The data processing module includes a preprocessing unit for preprocessing the collected operating data of the integrated distribution box to obtain optimized operating data, and a transmission unit connected to the preprocessing unit for transmitting the optimized operating data to the cloud platform via a 5G network. The analysis module, which is connected to the data processing module, includes a feature extraction unit for extracting features from the optimized running data to obtain running features, a model training unit connected to the feature extraction unit for constructing a machine learning model based on the running features, and an analysis unit connected to the model training unit for analyzing the running data based on the machine learning model to obtain analysis results. A remote control module, which is connected to the analysis module, is used to generate a control scheme based on the analysis results to remotely control the integrated distribution box; The control determination module, which is connected to the remote control module, is used to determine whether the remote control effectiveness of the integrated distribution box meets the requirements based on the accuracy of the analysis results and the control response time of the remote control. The trimming and adjustment module, which is connected to the control determination module, is used to determine the invalid representation trimming threshold of the machine learning model based on the redundancy rate of the operating characteristics when the remote control effectiveness of the integrated power distribution box does not meet the requirements. A constraint adjustment module, which is connected to the trimming adjustment module, is used to determine the temporal neighborhood coupling constraint coefficient of the running characteristics based on the transmission delay jitter amplitude of the optimized running data.

[0006] Furthermore, the control determination module determines the control characterization value based on the ratio of the accuracy of the analysis results to the control response time of the remote control, thereby determining whether the remote control effectiveness of the integrated distribution box meets the requirements.

[0007] Furthermore, the control determination module determines that the remote control effectiveness of the integrated distribution box meets the requirements when the control characterization value is greater than or equal to the preset control characterization value. The control determination module determines that the remote control effectiveness of the integrated distribution box does not meet the requirements when the control characterization value is less than the preset control characterization value.

[0008] Furthermore, in response to the condition that the remote control effectiveness of the integrated power distribution box does not meet the requirements, the trimming and adjustment module determines whether the extraction effectiveness of the operating features meets the requirements based on the redundancy rate of the operating features.

[0009] Furthermore, the trimming and adjustment module determines that the extraction validity of the operating feature meets the requirements when the redundancy rate of the operating feature is less than or equal to a preset first redundancy rate. The trimming and adjustment module determines that the extraction validity of the operating feature does not meet the requirements when the redundancy rate of the operating feature is greater than the preset first redundancy rate.

[0010] Furthermore, the trimming adjustment module increases the invalid representation trimming threshold of the machine learning model in response to the redundancy rate of the running feature being greater than the preset first redundancy rate and less than the preset second redundancy rate. The trimming and adjustment module responds to the fact that the redundancy rate of the operating feature is greater than or equal to the preset second redundancy rate, and initially determines that the transmission validity of the operating data does not meet the requirements. It then determines whether the transmission validity of the operating data meets the requirements based on the amplitude of the transmission delay jitter of the optimized operating data.

[0011] Furthermore, the increase in the invalid representation pruning threshold of the machine learning model is determined by the difference between the redundancy rate of the running features and the preset first redundancy rate.

[0012] Furthermore, in response to the condition that the redundancy rate of the operating feature is greater than or equal to the preset second redundancy rate, the constraint adjustment module determines whether the transmission validity of the operating data meets the requirements based on the amplitude of the transmission delay jitter of the optimized operating data.

[0013] Furthermore, the constraint adjustment module determines that the transmission validity of the operation data meets the requirements when the transmission delay jitter amplitude of the optimized operation data is less than or equal to the preset jitter amplitude. The constraint adjustment module responds when the transmission delay jitter of the optimized running data exceeds the preset jitter amplitude, determines that the transmission validity of the running data does not meet the requirements, and reduces the temporal neighborhood coupling constraint coefficient of the running features.

[0014] Furthermore, the reduction in the temporal neighborhood coupling constraint coefficient of the operating feature is determined by the difference between the transmission delay jitter amplitude of the optimized operating data and the preset jitter amplitude.

[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: The system of this invention, by setting up a data processing module, an analysis module, a remote control module, a control judgment module, a pruning and adjustment module, and a constraint adjustment module, determines the effectiveness of remote control of the integrated power distribution box based on the accuracy of the analysis results and the control representation value determined by the remote control response time. Due to the high dimensionality of the operational features, the training time of the machine learning model is increased, but the recognition effect and analysis capability are not significantly improved. Model training and optimization cannot be completed within the specified time, resulting in an excessively long training cycle and missing the model update window, thus preventing the machine learning model from being updated in a timely manner. By determining the effectiveness of remote control of the integrated power distribution box, problems such as delayed machine learning model updates and slow command issuance can be detected in a timely manner, optimizing feature extraction and model training, reasonably controlling data computation consumption, ensuring regular model updates, and adjusting the invalid representation pruning threshold of the machine learning model based on the redundancy rate of the operational features. Since data packets are prone to out-of-order issues during 5G transmission, the cloud cannot directly utilize... Feature extraction using out-of-order data requires the introduction of alignment compensation auxiliary features to re-align out-of-order data packets, thereby increasing redundant features and raising the dimensionality of operational features. By increasing the invalid representation pruning threshold of the machine learning model, the screening and removal of redundant representations and invalid auxiliary features derived during the re-alignment process can be strengthened, the output scale of redundant features can be compressed, and the excessive expansion of operational features can be effectively suppressed. The temporal neighborhood coupling constraint coefficient of operational features is adjusted according to the transmission delay jitter of optimized operational data. Due to 5G transmission delay jitter, the data stream formed after edge node preprocessing cannot be guaranteed to arrive at the cloud in the collection sequence during transmission, resulting in out-of-order data packets. The cloud needs to introduce auxiliary features for re-alignment processing. By reducing the temporal neighborhood coupling constraint coefficient of operational features, the strength of temporal sequence neighborhood association constraint can be relaxed, adapting to stable transmission scenarios under low latency jitter, reducing additional auxiliary features generated by temporal constraints, improving the feature redundancy problem caused by data transmission jitter, and improving the analysis accuracy of integrated power distribution box operational data.

[0016] Furthermore, the system described in this invention determines the effectiveness of remote control of the integrated distribution box by setting preset control characterization values. Due to the high dimension of the operating features, the training time of the machine learning model will be increased, but the recognition effect and analysis capability will not be significantly improved. The model training and optimization cannot be completed within the specified time, resulting in an excessively long training cycle and missing the model update window. Consequently, the machine learning model cannot be updated in a timely manner. By determining the effectiveness of remote control of the integrated distribution box, the problems of delayed machine learning model updates and slow command issuance can be detected in a timely manner, the feature extraction and model training links can be optimized, the data operation consumption can be reasonably controlled, and the model can be updated regularly, thereby further improving the accuracy of the analysis of the operating data of the integrated distribution box.

[0017] Furthermore, the system of the present invention adjusts the invalid representation pruning threshold of the machine learning model by setting a preset first redundancy rate and a preset second redundancy rate. Since data packets are prone to out-of-order phenomena during 5G transmission, the cloud cannot directly use out-of-order data for feature extraction. Alignment compensation auxiliary features need to be introduced to re-align out-of-order data packets, thereby increasing some redundant features and increasing the dimension of operating features. By increasing the invalid representation pruning threshold of the machine learning model, the screening and elimination of redundant representations and invalid auxiliary features derived during the re-alignment process can be strengthened, the output scale of redundant features can be compressed, and the excessive expansion of operating features can be effectively suppressed, thereby further improving the analysis accuracy of integrated distribution box operating data.

[0018] Furthermore, the system described in this invention adjusts the temporal neighborhood coupling constraint coefficient of the operating features by setting a preset jitter amplitude. Due to the jitter of 5G transmission latency, the data stream formed after edge node preprocessing cannot be guaranteed to arrive at the cloud in the collection sequence during transmission, resulting in out-of-order data packets. The cloud needs to introduce auxiliary features for realignment processing. By reducing the temporal neighborhood coupling constraint coefficient of the operating features, the strength of the temporal sequence neighborhood association constraint can be relaxed, adapting to stable transmission scenarios under low latency jitter, reducing the additional auxiliary features generated by temporal constraints, improving the feature redundancy problem caused by data transmission jitter, and further improving the analysis accuracy of integrated power distribution box operating data. Attached Figure Description

[0019] Figure 1 This is an overall structural block diagram of the integrated power distribution box remote control system based on 5G communication according to an embodiment of the present invention; Figure 2 This is a flowchart illustrating the process of determining the effectiveness of remote control of an integrated power distribution box using a 5G communication-based remote control system according to an embodiment of the present invention. Figure 3 This is a flowchart illustrating the process of determining the invalid representation pruning threshold of a machine learning model in a remote control system for an integrated distribution box based on 5G communication, according to an embodiment of the present invention. Figure 4 This is a logical flowchart illustrating the process of determining the temporal neighborhood coupling constraint coefficients of the integrated distribution box remote control system based on 5G communication in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0021] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0022] Please see Figure 1 As shown, it is an overall structural block diagram of the integrated power distribution box remote control system based on 5G communication according to an embodiment of the present invention.

[0023] An embodiment of the present invention provides a remote control system for integrated distribution boxes based on 5G communication, comprising: The data processing module includes a preprocessing unit for preprocessing the collected operational data of the integrated distribution box through an edge computing node to obtain optimized operational data, and a transmission unit connected to the preprocessing unit for transmitting the optimized operational data to the cloud platform through a 5G network. The analysis module, which is connected to the data processing module, includes a feature extraction unit for extracting features from the optimized running data to obtain running features, a model training unit connected to the feature extraction unit for constructing a machine learning model based on the running features, and an analysis unit connected to the model training unit for analyzing the running data based on the machine learning model to obtain analysis results. A remote control module, which is connected to the analysis module, is used to generate a control scheme based on the analysis results to remotely control the integrated distribution box; The control determination module, which is connected to the remote control module, is used to determine whether the remote control effectiveness of the integrated distribution box meets the requirements based on the accuracy of the analysis results and the control response time of the remote control. The trimming and adjustment module, which is connected to the control determination module, is used to determine the invalid representation trimming threshold of the machine learning model based on the redundancy rate of the operating characteristics when the remote control effectiveness of the integrated power distribution box does not meet the requirements. A constraint adjustment module, which is connected to the trimming adjustment module, is used to determine the temporal neighborhood coupling constraint coefficient of the running characteristics based on the transmission delay jitter amplitude of the optimized running data.

[0024] Specifically, the operating data includes the incoming voltage of the integrated distribution box, the outgoing current of the integrated distribution box, and the ambient temperature inside the integrated distribution box.

[0025] Specifically, preprocessing includes data cleaning, noise reduction, and outlier removal.

[0026] Specifically, optimizing operational data includes the integrated distribution box's incoming line voltage after data cleaning, the integrated distribution box's outgoing line current after noise reduction, and the integrated distribution box's internal ambient temperature after data format standardization.

[0027] Specifically, the operating characteristics include the average incoming voltage of the integrated distribution box, the variance of the outgoing current of the integrated distribution box, and the rate of change of ambient temperature inside the integrated distribution box.

[0028] Specifically, the process of building a machine learning model based on operational features involves dividing the operational features into a training set and a test set, using the training set to train the initial model, using the validation set to optimize and adjust the parameters, and using the test set to evaluate the quality of the generated model, ultimately obtaining the machine learning model.

[0029] Specifically, the initial model is an initial framework that has basic power distribution operation data feature extraction capabilities and power distribution scenario state perception capabilities, and is suitable for 5G IoT power distribution data acquisition and edge computing preprocessing analysis.

[0030] Specifically, the machine learning model can be a gradient boosting tree model, a random forest model, or a temporal convolutional network, with the preferred embodiment being the gradient boosting tree model.

[0031] Specifically, the process of analyzing operational data based on machine learning models to obtain analysis results involves extracting features from the collected and edge computing preprocessed operational data, outputting operational features, performing pattern recognition and quantitative evaluation on the operational features based on machine learning models, and finally outputting analysis results.

[0032] Specifically, the process of generating control schemes based on analysis results to remotely control the integrated distribution box involves identifying abnormal operation problems of the integrated distribution box based on the analysis results, matching the corresponding control rules to generate control schemes, converting the control schemes into control commands, and then sending them to the integrated distribution box via the 5G network to achieve remote control.

[0033] Specifically, the analysis results include fault diagnosis results such as short circuit, excessive harmonics, and overload prediction.

[0034] Specifically, the control scheme includes circuit breaker tripping, activating active filters, and issuing power limiting commands to controllable loads.

[0035] Specifically, the accuracy rate of the analysis results is the ratio of the number of accurate analysis results to the total number of analysis results.

[0036] Specifically, accuracy means that the analysis results are consistent with the actual electrical operating status of the integrated distribution box.

[0037] Specifically, the redundancy rate of an operational feature is the ratio of the number of redundant features to the total number of operational features.

[0038] Specifically, redundancy is characterized by ineffectiveness, duplication, or additional compensation, and does not possess information about the actual operating conditions of the distribution box.

[0039] Specifically, the invalid representation pruning threshold of a machine learning model is a critical value used to identify and prune invalid representations, reducing feature dimension redundancy.

[0040] Specifically, the jitter amplitude of the optimized running data transmission delay is the difference between the maximum and minimum transmission delay times of the optimized running data.

[0041] Specifically, latency jitter refers to the fluctuation in the transmission time of optimized operational data during the process of transmitting the data from the edge node to the cloud platform via the 5G network.

[0042] Specifically, latency is the time consumed for optimized operational data to be transmitted from edge nodes to the cloud platform via the 5G network.

[0043] Specifically, the temporal neighborhood coupling constraint coefficient of the running feature is used to control the degree of correlation between running features at adjacent time points.

[0044] In implementation, the system of this invention sets up a data processing module, an analysis module, a remote control module, a control judgment module, a pruning and adjustment module, and a constraint adjustment module. It judges the effectiveness of remote control of the integrated power distribution box based on the accuracy of the analysis results and the control representation value determined by the remote control response time. Due to the high dimensionality of the operational features, the training time of the machine learning model increases, but the recognition effect and analysis capability are not significantly improved. Model training and optimization cannot be completed within the specified time, resulting in an excessively long training cycle and missing the model update window, thus preventing the machine learning model from being updated in a timely manner. By judging the effectiveness of remote control of the integrated power distribution box, problems such as delayed machine learning model updates and slow command issuance can be detected in a timely manner. Feature extraction and model training can be optimized, data processing consumption can be reasonably controlled, and regular model updates can be ensured. The invalid representation pruning threshold of the machine learning model is adjusted according to the redundancy rate of the operational features. Since data packets are prone to out-of-order phenomena during 5G transmission, the cloud cannot directly utilize out-of-order data. Feature extraction requires the introduction of alignment compensation auxiliary features to re-align out-of-order data packets, which increases some redundant features and raises the dimensionality of the operational features. By increasing the invalid representation pruning threshold of the machine learning model, the screening and removal of redundant representations and invalid auxiliary features derived during the re-alignment process can be strengthened, the output scale of redundant features can be compressed, and the excessive expansion of operational features can be effectively suppressed. The temporal neighborhood coupling constraint coefficient of the operational features is adjusted according to the transmission delay jitter of the optimized operational data. Due to the 5G transmission delay jitter, the data stream formed after edge node preprocessing cannot be guaranteed to arrive at the cloud in the collection sequence during transmission, resulting in out-of-order data packets. The cloud needs to introduce auxiliary features for re-alignment processing. By reducing the temporal neighborhood coupling constraint coefficient of the operational features, the strength of the temporal sequence neighborhood association constraint can be relaxed, adapting to stable transmission scenarios under low latency jitter, reducing the additional auxiliary features generated by temporal constraints, improving the feature redundancy problem caused by data transmission jitter, and improving the analysis accuracy of integrated distribution box operation data.

[0045] Please continue reading. Figure 2 As shown, it is a logical flowchart of the process of determining whether the remote control effectiveness of the integrated power distribution box meets the requirements of the 5G communication-based integrated power distribution box remote control system according to an embodiment of the present invention.

[0046] Specifically, the control determination module determines the control characterization value by the ratio of the accuracy of the analysis results to the control response time of the remote control, in order to determine whether the remote control effectiveness of the integrated distribution box meets the requirements.

[0047] Specifically, the control determination module determines that the remote control effectiveness of the integrated power distribution box meets the requirements when the control characterization value is greater than or equal to the preset control characterization value. The control determination module determines that the remote control effectiveness of the integrated distribution box does not meet the requirements when the control characterization value is less than the preset control characterization value.

[0048] Understandably, in a 5G-based integrated distribution box remote control system, the core logic of using preset control characterization values ​​to characterize the effectiveness of remote control is to quantify the accuracy of distribution box operating condition analysis and the 5G command transmission response efficiency into quantifiable evaluation indicators. The control characterization value is calculated by analyzing the accuracy of the results and the remote control response time, and then compared with the preset control characterization value to achieve a comprehensive judgment on the 5G data transmission quality, the accuracy of operating feature extraction, the machine learning model analysis capability, and the remote command execution efficiency of the equipment. When the control performance value is greater than or equal to the preset control performance value, it indicates that the timing alignment of the power distribution operation data is good, the interference of invalid redundancy features is weak, the model operating condition analysis results are accurate and reliable, the 5G communication link transmission is stable, the remote control command issuance and execution response are timely, the overall remote control operation of the integrated distribution box is stable, and the control effectiveness meets the actual use needs of remote monitoring, intelligent control, and safe operation and maintenance in the power distribution field. When the control performance value is less than the preset control performance value, it indicates that the 5G transmission latency jitter and data packet out-of-order problems are prominent, the redundancy of operating features is high, the deviation of the model analysis results is increased, the remote control command transmission delay is high, the execution feedback is lagging, the remote linkage control capability of the distribution box is reduced, the overall control stability and real-time performance are insufficient, and the remote control effectiveness does not meet the standard. It is necessary to correct the invalid performance pruning threshold through the pruning and adjustment module and optimize the timing neighborhood coupling constraint coefficient through the constraint adjustment module to complete the adaptive optimization of model parameters and timing constraint rules. The preset control performance value can be set according to the actual operating conditions. The preset control performance value aims to ensure the effectiveness and practicality of remote control of the integrated distribution box. Optionally, the preset control characterization value is determined through a limited number of tests by evaluating the remote control effect of different control characterization values ​​on the integrated distribution box. The determined preset control characterization value should be neither too small nor cause excessive interference to the remote control process of the integrated distribution box. For example, the preset control characterization value is generally selected in the range of [0.5% / ms, 1% / ms].

[0049] Preferably, the preferred embodiment of the preset control characterization value is 0.7% / ms.

[0050] Specifically, % / ms is the unit of the manipulation representation value, meaning a percentage per millisecond.

[0051] In practice, the system described in this invention determines the effectiveness of remote control of the integrated distribution box by setting preset control characterization values. Due to the high dimension of the operating features, the training time of the machine learning model is increased, but the recognition effect and analysis capability are not significantly improved. The model training and optimization cannot be completed within the specified time, resulting in an excessively long training cycle and missing the model update window. Consequently, the machine learning model cannot be updated in a timely manner. By determining the effectiveness of remote control of the integrated distribution box, the problems of delayed machine learning model updates and slow command issuance can be detected in a timely manner. The feature extraction and model training links can be optimized, data operation consumption can be reasonably controlled, and the model can be updated regularly, further improving the accuracy of the analysis of the integrated distribution box operation data.

[0052] Please continue reading. Figure 3 As shown, it is a logical flowchart of the process of determining the invalid representation pruning threshold of the machine learning model in the remote control system of integrated distribution box based on 5G communication according to an embodiment of the present invention.

[0053] Specifically, in response to the condition that the remote control effectiveness of the integrated power distribution box does not meet the requirements, the trimming and adjustment module determines whether the extraction effectiveness of the operating features meets the requirements based on the redundancy rate of the operating features.

[0054] Specifically, the trimming and adjustment module determines that the extraction validity of the running feature meets the requirements when the redundancy rate of the running feature is less than or equal to a preset first redundancy rate. The trimming and adjustment module determines that the extraction validity of the operating feature does not meet the requirements when the redundancy rate of the operating feature is greater than the preset first redundancy rate.

[0055] Specifically, the trimming adjustment module responds to the fact that the redundancy rate of the running feature is greater than the preset first redundancy rate and less than the preset second redundancy rate by increasing the invalid representation trimming threshold of the machine learning model. The trimming and adjustment module responds to the fact that the redundancy rate of the operating feature is greater than or equal to the preset second redundancy rate, and initially determines that the transmission validity of the operating data does not meet the requirements. It then determines whether the transmission validity of the operating data meets the requirements based on the amplitude of the transmission delay jitter of the optimized operating data.

[0056] It is understandable that the preset first redundancy rate is less than the preset second redundancy rate, and the three intervals divided by the preset first redundancy rate and the preset second redundancy rate correspond to three different scenarios: The first interval is when the redundancy rate of the running feature is less than or equal to the preset first redundancy rate. The corresponding situation is: the extraction validity of the running feature is determined to meet the requirements. The second interval is when the redundancy rate of the running feature is greater than the preset first redundancy rate and less than the preset second redundancy rate. The corresponding situation is: because data packets are prone to out-of-order phenomena during 5G transmission, the cloud cannot directly use out-of-order data for feature extraction. Alignment compensation auxiliary features need to be introduced to re-align out-of-order data packets, thereby increasing some redundant features and causing the dimension of the running feature to increase. The third interval is when the redundancy rate of the operating feature is greater than or equal to the preset second redundancy rate. The corresponding situation is: due to the jitter of 5G transmission latency, the data stream formed after edge node preprocessing cannot be guaranteed to arrive at the cloud in the collection sequence during transmission, resulting in data packets being out of order. The cloud needs to introduce auxiliary features for realignment processing.

[0057] Understandably, in a 5G-based integrated distribution box remote control system, the introduction of preset first redundancy rate and preset second redundancy rate characterizes the extraction quality and redundancy control level of operational features. The core logic is to transform the effectiveness of operational feature extraction into a quantifiable redundancy rate range. The preset first redundancy rate is the boundary between qualified operational feature extraction and the need to increase the threshold for invalid feature pruning. The preset second redundancy rate is the critical point for determining whether the pruning threshold needs adjustment and whether the operational data transmission effectiveness is abnormal. This provides a quantitative basis for targeted identification of issues caused by feature extraction redundancy, 5G transmission disorder, and time delay jitter. The preset first and second redundancy rates can be set according to actual operating conditions. The preset first and second redundancy rates aim to ensure the effectiveness and practicality of remote control of the integrated distribution box. Optionally, the preset first and second redundancy rates are determined through a limited number of tests by evaluating the effect of different redundancy rates on the remote control of the integrated distribution box. The determined preset first and second redundancy rates should be neither too small nor cause excessive interference to the remote control process of the integrated distribution box. For example, the preset first redundancy rate is generally selected in the range of [6%, 12%], and the preset second redundancy rate is generally selected in the range of [13%, 20%].

[0058] Preferably, the first redundancy rate is 9% in a preferred embodiment, and the second redundancy rate is 16% in a preferred embodiment.

[0059] Specifically, the increase in the invalid representation pruning threshold of the machine learning model is determined by the difference between the redundancy rate of the running features and a preset first redundancy rate.

[0060] Specifically, when the difference between the redundancy rate of the running features and the preset first redundancy rate is within 5%, the invalid representation pruning threshold of the machine learning model is increased to 1.08 times the original value. When the difference between the redundancy rate of the running features and the preset first redundancy rate exceeds 5%, in addition to increasing to 1.08 times the original value, for every 1% exceeding 5%, the invalid representation pruning threshold of the machine learning model increases by 0.005. For example, if the difference between the redundancy rate of the running features and the preset first redundancy rate is 7%, and the current invalid representation pruning threshold of the machine learning model is 0.3, then the increased invalid representation pruning threshold of the machine learning model is 0.3 × 1.08 + 0.005 × 2 = 0.334.

[0061] In implementation, the system of the present invention adjusts the invalid representation pruning threshold of the machine learning model by setting a preset first redundancy rate and a preset second redundancy rate. Since data packets are prone to out-of-order phenomena during 5G transmission, the cloud cannot directly use out-of-order data for feature extraction. Alignment compensation auxiliary features need to be introduced to re-align out-of-order data packets, thereby increasing some redundant features and increasing the dimension of operating features. By increasing the invalid representation pruning threshold of the machine learning model, the screening and elimination of redundant representations and invalid auxiliary features derived during the re-alignment process can be strengthened, the scale of redundant feature output can be compressed, the excessive expansion of operating features can be effectively suppressed, and the analysis accuracy of integrated distribution box operation data can be further improved.

[0062] Please continue reading. Figure 4 As shown, it is a logical flowchart of the process of determining the temporal neighborhood coupling constraint coefficient of the integrated distribution box remote control system based on 5G communication in an embodiment of the present invention.

[0063] Specifically, in response to the condition that the redundancy rate of the operating feature is greater than or equal to the preset second redundancy rate, the constraint adjustment module determines whether the transmission validity of the operating data meets the requirements based on the amplitude of the transmission delay jitter of the optimized operating data.

[0064] Specifically, the constraint adjustment module determines that the transmission validity of the operation data meets the requirements when the transmission delay jitter amplitude of the optimized operation data is less than or equal to the preset jitter amplitude. The constraint adjustment module responds when the transmission delay jitter of the optimized running data exceeds the preset jitter amplitude, determines that the transmission validity of the running data does not meet the requirements, and reduces the temporal neighborhood coupling constraint coefficient of the running features.

[0065] It is understandable that the two preset intervals for the jitter amplitude correspond to two different scenarios: The first interval is when the jitter amplitude of the transmission delay of the optimized running data is less than or equal to the preset jitter amplitude. The corresponding situation is: the validity of the transmission of the running data is determined to meet the requirements. The second interval is when the transmission latency jitter of the optimized running data is greater than the preset jitter amplitude. The corresponding situation is: due to the 5G transmission latency jitter, the data stream formed after edge node preprocessing cannot be guaranteed to arrive at the cloud in the collection sequence during transmission, resulting in out-of-order data packets. The cloud needs to introduce auxiliary features for realignment processing.

[0066] Understandably, using a preset jitter amplitude to characterize the transmission effectiveness of operational data involves transforming the timing stability of 5G remote transmission into a quantifiable amplitude metric. By comparing the optimized transmission delay jitter amplitude with the preset jitter amplitude, the strength of timing disturbances in the communication link and the severity of data packet out-of-order issues can be accurately identified. This provides a quantitative basis for adaptive adjustment of timing constraint parameters, feature extraction optimization, and remote control stability control. Transmission delay jitter amplitude is a core indicator for measuring data transmission quality. Excessive jitter will directly disrupt the data flow timing structure, continuously increasing the pressure on cloud realignment processing, leading to increased model feature parsing load and increased remote command latency, severely impacting the real-time performance and accuracy of remote monitoring and fault diagnosis of the integrated distribution box. The preset jitter amplitude can be set according to actual operating conditions. The preset jitter amplitude setting aims to ensure the effectiveness and practicality of remote control of the integrated distribution box. Optionally, the preset jitter amplitude is determined through a limited number of tests by evaluating the effect of different transmission delay jitter amplitudes of optimized operating data on the remote control of the integrated distribution box. The determined preset jitter amplitude should be neither too small nor cause excessive interference to the remote control process of the integrated distribution box. For example, the preset jitter amplitude is generally selected in the range of [8ms, 15ms].

[0067] Preferably, the preset jitter amplitude is 12ms in a preferred embodiment.

[0068] Specifically, the reduction in the temporal neighborhood coupling constraint coefficient of the operating feature is determined by the difference between the transmission delay jitter amplitude of the optimized operating data and the preset jitter amplitude.

[0069] Specifically, when the difference between the transmission delay jitter of the optimized running data and the preset jitter is within 5ms, the temporal neighborhood coupling constraint coefficient of the running feature is reduced to 0.98 times its original value. When the difference between the transmission delay jitter of the optimized running data and the preset jitter exceeds 5ms, in addition to reducing it to 0.98 times its original value, the temporal neighborhood coupling constraint coefficient of the running feature is reduced by 0.002 for every 1ms exceeding 5ms. For example, if the difference between the transmission delay jitter of the optimized running data and the preset jitter is 8ms, and the current temporal neighborhood coupling constraint coefficient of the running feature is 0.6, the reduced temporal neighborhood coupling constraint coefficient of the running feature is 0.6×0.98-0.002×3=0.582.

[0070] In implementation, the system of the present invention adjusts the temporal neighborhood coupling constraint coefficient of the operating features by setting a preset jitter amplitude. Due to the jitter of 5G transmission latency, the data stream formed after edge node preprocessing cannot be guaranteed to arrive at the cloud in the collection sequence during transmission, resulting in out-of-order data packets. The cloud needs to introduce auxiliary features for realignment processing. By reducing the temporal neighborhood coupling constraint coefficient of the operating features, the neighborhood association constraint strength of the temporal sequence can be relaxed, adapting to stable transmission scenarios under low latency jitter, reducing the additional auxiliary features generated by the temporal constraints, improving the feature redundancy problem caused by data transmission jitter, and further improving the analysis accuracy of the integrated power distribution box operating data.

[0071] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A remote control system for integrated distribution boxes based on 5G communication, characterized in that, include: The data processing module includes a preprocessing unit for preprocessing the collected operating data of the integrated distribution box to obtain optimized operating data, and a transmission unit connected to the preprocessing unit for transmitting the optimized operating data to the cloud platform via a 5G network. The analysis module, which is connected to the data processing module, includes a feature extraction unit for extracting features from the optimized running data to obtain running features, a model training unit connected to the feature extraction unit for constructing a machine learning model based on the running features, and an analysis unit connected to the model training unit for analyzing the running data based on the machine learning model to obtain analysis results. A remote control module, which is connected to the analysis module, is used to generate a control scheme based on the analysis results to remotely control the integrated distribution box; The control determination module, which is connected to the remote control module, is used to determine whether the remote control effectiveness of the integrated distribution box meets the requirements based on the accuracy of the analysis results and the control response time of the remote control. The trimming and adjustment module, which is connected to the control determination module, is used to determine the invalid representation trimming threshold of the machine learning model based on the redundancy rate of the operating characteristics when the remote control effectiveness of the integrated power distribution box does not meet the requirements. A constraint adjustment module, which is connected to the trimming adjustment module, is used to determine the temporal neighborhood coupling constraint coefficient of the running characteristics based on the transmission delay jitter amplitude of the optimized running data. The control determination module responds to the ratio of the accuracy of the analysis results to the control response time of the remote control, which determines the control characterization value to determine whether the remote control effectiveness of the integrated distribution box meets the requirements. The trimming adjustment module increases the invalid representation trimming threshold of the machine learning model in response to the redundancy rate of the running feature being greater than a preset first redundancy rate and less than a preset second redundancy rate. The trimming and adjustment module responds to the fact that the redundancy rate of the running feature is greater than or equal to the preset second redundancy rate, and initially determines that the transmission validity of the running data does not meet the requirements. It then determines whether the transmission validity of the running data meets the requirements based on the jitter amplitude of the transmission delay of the optimized running data. The constraint adjustment module responds when the transmission delay jitter of the optimized running data exceeds a preset jitter amplitude, determines that the transmission validity of the running data does not meet the requirements, and reduces the temporal neighborhood coupling constraint coefficient of the running features. The redundancy rate of a running feature is the ratio of the number of redundant features to the total number of running features. The invalid representation pruning threshold of a machine learning model is a critical value used to identify and prune invalid representations, reducing feature dimension redundancy. The temporal neighborhood coupling constraint coefficient of the running feature is used to control the degree of correlation between running features at adjacent time points.

2. The integrated power distribution box remote control system based on 5G communication according to claim 1, characterized in that, The control determination module determines that the remote control effectiveness of the integrated distribution box meets the requirements when the control characterization value is greater than or equal to the preset control characterization value. The control determination module determines that the remote control effectiveness of the integrated distribution box does not meet the requirements when the control characterization value is less than the preset control characterization value.

3. The integrated distribution box remote control system based on 5G communication according to claim 2, characterized in that, When the remote control effectiveness of the integrated power distribution box does not meet the requirements, the trimming and adjustment module determines whether the extraction effectiveness of the operating features meets the requirements based on the redundancy rate of the operating features.

4. The integrated distribution box remote control system based on 5G communication according to claim 3, characterized in that, The trimming and adjustment module determines that the extraction validity of the running feature meets the requirements when the redundancy rate of the running feature is less than or equal to the preset first redundancy rate. The trimming and adjustment module determines that the extraction validity of the operating feature does not meet the requirements when the redundancy rate of the operating feature is greater than the preset first redundancy rate.

5. The integrated distribution box remote control system based on 5G communication according to claim 4, characterized in that, The increase in the invalid representation pruning threshold of the machine learning model is determined by the difference between the redundancy rate of the running features and the preset first redundancy rate.

6. The integrated distribution box remote control system based on 5G communication according to claim 5, characterized in that, The constraint adjustment module, in response to the condition that the redundancy rate of the operating feature is greater than or equal to the preset second redundancy rate, determines whether the transmission validity of the operating data meets the requirements based on the amplitude of the transmission delay jitter of the optimized operating data.

7. The integrated distribution box remote control system based on 5G communication according to claim 6, characterized in that, The constraint adjustment module determines that the transmission validity of the operation data meets the requirements when the transmission delay jitter of the optimized operation data is less than or equal to the preset jitter amplitude.

8. The integrated distribution box remote control system based on 5G communication according to claim 7, characterized in that, The reduction in the temporal neighborhood coupling constraint coefficient of the operating feature is determined by the difference between the transmission delay jitter amplitude of the optimized operating data and the preset jitter amplitude.

Images