An intelligent power distribution cabinet with fault detection function and a detection method

Abstract

The present application relates to the technical field of smart grid, especially to a kind of intelligent power distribution cabinet with fault detection function and detection method.The present application can accurately obtain state data and analyze state data by obtaining unit and data analysis unit;Multiple AI analysis models in edge computing unit can accurately analyze equipment state results, and can improve the efficiency of power distribution cabinet fault detection;Through the decision unit, multiple equipment state results are fused using the preset rules to obtain accurate fault information, and the execution module in the control unit is driven based on the accurate fault information to act;Finally, the analysis unit determines the fault detection state based on the fault suppression success rate within the preset period, and adjusts the corresponding parameters based on the fault detection state, so as to more accurately analyze the fault.The present application improves the fault detection efficiency of power distribution cabinet.

Description

Technical Field

[0001] This invention relates to the field of smart grid technology, and in particular to a smart distribution cabinet with fault detection function and a detection method. Background Technology

[0002] Distribution cabinets are critical equipment in power transmission and distribution systems, and their operational status directly affects the reliability and security of power supply. Traditional distribution cabinet fault detection relies heavily on manual periodic inspections, simple instrument readings (such as voltmeters and ammeters), or reactive protection mechanisms like circuit breaker tripping. This approach has the following drawbacks: insufficient early warning capability (it cannot provide early warnings for latent faults such as overheating, insulation aging, and poor contact, often only discovering them after the fault has occurred or escalated); poor real-time performance (manual inspections are time-consuming, making 24 / 7 monitoring impossible and hindering timely detection of sudden faults); low diagnostic accuracy (relying on the experience of maintenance personnel, lacking data support, and prone to misdiagnosis or omission); and isolated data (the lack of effective digital archives hinders fault tracing and lifecycle management).

[0003] Chinese Patent Publication No. CN119337295A discloses a method and system for detecting faults in power distribution cabinets. The method includes deploying a multimodal sensor array, constructing a multi-dimensional data acquisition network, collecting various types of sensor data in real time and performing data preprocessing; establishing a multimodal data fusion model using deep learning algorithms, forming a fault feature mapping relationship through feature extraction, data correlation analysis, and cross-validation to achieve collaborative perception at the data level; and constructing a multi-dimensional fault diagnosis decision tree based on the collaborative perception results, combining the weight allocation and confidence assessment of each sensor data to accurately locate the specific device where the fault occurred, and providing an assessment result of the fault type and severity.

[0004] It is evident that existing technologies suffer from the following problems: traditional power distribution cabinets exhibit delays, limitations, dependence, and isolation in fault detection, resulting in low fault detection efficiency. Summary of the Invention

[0005] Therefore, the present invention provides an intelligent power distribution cabinet and a detection method with fault detection function to overcome the shortcomings of traditional power distribution cabinets in fault detection, such as lag, one-sidedness, dependence and isolation, which result in low fault detection efficiency of power distribution cabinets.

[0006] To achieve the above objectives, the present invention provides an intelligent power distribution cabinet with fault detection function, comprising:

[0007] The acquisition unit uses a multi-sensor array installed in the distribution cabinet to collect multi-dimensional status data, including at least temperature data of the heat concentration area, high-frequency harmonic signals of line current and environmental parameters inside the distribution cabinet.

[0008] A data parsing unit, which is connected to the acquisition unit, is used to parse the multi-dimensional state data;

[0009] An edge computing unit, connected to the data parsing unit, is used to run at least one AI analysis model on the parsed multi-dimensional state data and output at least one device state analysis result corresponding to the multi-dimensional state data.

[0010] A decision unit, connected to the edge computing unit, is used to perform information fusion on the status analysis results of at least one of the devices based on preset rules, so as to output fault information, including fault type and fault location.

[0011] A control unit, which is connected to the decision unit, is used to control the execution module in the power distribution cabinet to perform actions based on the fault information obtained through communication transmission;

[0012] An analysis unit, connected to the control unit, calculates the fault suppression success rate within a preset period, determines the fault detection status based on the fault suppression success rate, adjusts the transmission rate between the decision unit and the control unit based on the fault detection status, and adjusts the startup cycle of the model incremental learning module based on the fault detection status after adjusting the transmission rate. The fault suppression success rate is the ratio of the number of times a potential fault is eliminated to the number of actions of the execution module. The model incremental learning module is a module in the edge computing unit that optimizes the AI ​​model using newly generated data.

[0013] Furthermore, the analysis unit is also used to calculate the difference between a preset value and the fault suppression success rate within a preset period when the fault detection state is determined to be unqualified; the analysis unit is also used to adjust the transmission rate between the decision unit and the control unit based on the ratio of the difference to the preset difference when the difference is greater than the preset difference; wherein, the fault detection state is unqualified when the fault suppression success rate is less than the preset value.

[0014] Furthermore, the analysis unit is also used to increase the transmission rate between the decision unit and the control unit based on the ratio of the difference to a preset difference, and the increase in transmission rate is proportional to the ratio.

[0015] Furthermore, the decision-making unit also includes a rule-establishing module, which is used to establish different rule parameter sets according to different time periods to construct the preset rules; the analysis unit also includes obtaining the fault suppression success rate of multiple historical periods when the fault detection status is unqualified after adjusting the transmission rate; the analysis unit is also used to use the autocorrelation function to determine whether the plotted historical period-fault suppression success rate curve has periodicity; the decision-making unit is also used to activate the rule-establishing module when the curve has periodicity.

[0016] Furthermore, the analysis unit is also used to calculate the integral of the historical cycle-fault suppression success rate curve when the fault detection status is unqualified after the rule establishment module is started; the analysis unit is also used to adjust the start cycle of the model incremental learning module based on the ratio of the preset integral to the integral when the integral is less than the preset integral.

[0017] Furthermore, the analysis unit is also used to reduce the startup cycle of the model incremental learning module based on the ratio of a preset integral to the integral, and the reduction in startup cycle is proportional to the ratio.

[0018] Furthermore, the analysis unit is also used to, if the fault detection status is unqualified after adjusting the startup cycle of the model incremental learning module, repeatedly adjust the startup cycle of the model incremental learning module at least once until the number of adjustments is less than a preset number and the fault detection status is qualified, or the number of adjustments is equal to the preset number, and then stop adjusting; the analysis unit is also used to, if the fault detection status is unqualified after stopping adjustment, calculate the variance of the fault suppression success rate of multiple historical cycles; the analysis unit is also used to, if the variance is less than a preset variance, adjust the number of different sensors in the multi-sensor array based on the difference between the preset variance and the variance.

[0019] Furthermore, the analysis unit is also used to increase the number of different sensors in the multi-sensor array based on the difference between the preset variance and the variance, and the increase in the number of different sensors in the multi-sensor array is proportional to the difference.

[0020] Furthermore, the analysis unit is also used to reduce the sensor verification cycle based on the difference between the preset number of sensors in the multi-sensor array and the actual number of sensors, and the reduction in the verification cycle is inversely proportional to the difference.

[0021] To achieve the above objectives, the present invention provides a detection method applied to an intelligent power distribution cabinet with fault detection function as described in any one of the above claims, comprising:

[0022] The acquisition unit uses a multi-sensor array installed in the distribution cabinet to collect multi-dimensional status data, including at least temperature data of the heat concentration area, high-frequency harmonic signals of line current and environmental parameters inside the distribution cabinet.

[0023] The multi-dimensional state data is analyzed by a data parsing unit connected to the acquisition unit;

[0024] The edge computing unit connected to the data parsing unit runs at least one AI analysis model on the parsed multi-dimensional state data and outputs at least one device state analysis result corresponding to the multi-dimensional state data.

[0025] The decision unit connected to the edge computing unit performs information fusion on the status analysis results of at least one of the devices based on preset rules to output fault information, including fault type and fault location.

[0026] The control unit connected to the decision unit controls the execution module in the power distribution cabinet to perform actions based on the fault information obtained through communication transmission.

[0027] The analysis unit connected to the control unit calculates the fault suppression success rate within a preset period, determines the fault detection status based on the fault suppression success rate, adjusts the transmission rate between the decision unit and the control unit based on the fault detection status, and adjusts the startup cycle of the model incremental learning module based on the fault detection status after adjusting the transmission rate. The fault suppression success rate is the ratio of the number of times the potential fault is eliminated to the number of actions of the execution module. The model incremental learning module is a module in the edge computing unit that optimizes the AI ​​model using newly generated data.

[0028] Compared with existing technologies, the advantages of this invention are as follows: Through the acquisition unit and data parsing unit, it can accurately acquire and parse state data; through multiple AI analysis models in the edge computing unit, it can accurately analyze equipment state results and improve the efficiency of power distribution cabinet fault detection; through the decision unit, it can fuse multiple equipment state results using preset rules to obtain accurate fault information, and drive the execution module in the control unit to perform actions based on this accurate fault information; finally, through the analysis unit, it determines the fault detection status based on the fault suppression success rate within a preset period, and adjusts the corresponding parameters based on the fault detection status, thereby analyzing faults more accurately. This invention improves the fault detection efficiency of power distribution cabinets.

[0029] Furthermore, when the fault detection status is determined to be unqualified, the present invention determines the adjustment of corresponding parameters based on the difference between the preset value and the fault suppression success rate within a preset period. This allows for more accurate and effective parameter adjustment based on more precise causes, thereby solving the problem of one-sidedness in fault detection of power distribution cabinets and further improving the fault detection efficiency of power distribution cabinets.

[0030] Furthermore, this invention increases the transmission rate between the decision unit and the control unit based on the ratio of the difference to the preset difference, which can more accurately adjust the transmission rate between the decision unit and the control unit, enabling the control unit to act more promptly, thereby improving the efficiency of the power distribution cabinet after fault detection, and further improving the fault detection efficiency of the power distribution cabinet.

[0031] Furthermore, by determining whether the drawn historical cycle-fault suppression success rate curve has periodicity, the present invention determines whether to activate the rule establishment module, and can adaptively adjust the preset rules according to the current situation, thereby further improving the fault detection efficiency of the power distribution cabinet.

[0032] Furthermore, this invention determines whether to adjust the startup cycle of the incremental learning module of the model based on the integral of the historical cycle-fault suppression success rate curve, which can more effectively adjust based on the reasons for the failure of the fault detection status, thereby further improving the fault detection efficiency of the distribution cabinet.

[0033] Furthermore, this invention adjusts the startup cycle of the incremental learning module of the model based on the ratio of preset integrals, which can optimize the AI ​​analysis model more promptly based on newly generated data, thereby further improving the fault detection efficiency of the power distribution cabinet.

[0034] Furthermore, this invention determines whether to adjust the number of different sensors in the multi-sensor array based on the variance of the fault suppression success rate at multiple historical moments. This allows for more effective adjustment based on the reasons for the fault detection status being unqualified, thereby further improving the fault detection efficiency of the power distribution cabinet.

[0035] Furthermore, this invention adjusts the number of different sensors in a multi-sensor array based on a preset variance and the difference between the variances. By deploying an accurate number of sensors, outliers can be more effectively eliminated during data verification, thereby making the acquired multi-dimensional status data more accurate and further improving the fault detection efficiency of the power distribution cabinet.

[0036] Furthermore, this invention adjusts the sensor verification cycle based on the difference between the preset number of sensors and the actual number of sensors in the multi-sensor array. This allows for more effective checking of the sensor health status by adjusting the sensor verification cycle, while ensuring the overall reliability of the data provided by the multi-sensor array. This results in more accurate multi-dimensional status data and further improves the fault detection efficiency of the power distribution cabinet. Attached Figure Description

[0037] Figure 1 This is a schematic diagram of the structure of an intelligent power distribution cabinet with fault detection function according to an embodiment of the present invention;

[0038] Figure 2This is a flowchart illustrating the steps of a detection method for an intelligent power distribution cabinet with fault detection function, as described in an embodiment of the present invention.

[0039] Figure 3 This is a flowchart illustrating the steps of determining the success rate of fault suppression based on the comparison with a preset value in an embodiment of the present invention.

[0040] Figure 4 This is a flowchart illustrating the steps for determining the fault detection status based on adjusting the startup cycle of the incremental learning module in an embodiment of the present invention. Detailed Implementation

[0041] 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.

[0042] 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.

[0043] It should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0044] Please see Figure 1 As shown, it is a structural schematic diagram of an intelligent power distribution cabinet with fault detection function according to an embodiment of the present invention.

[0045] The system includes an acquisition unit, a data parsing unit, an edge computing unit, a decision-making unit, a control unit, and an analysis unit.

[0046] The acquisition unit uses a multi-sensor array installed in the distribution cabinet to collect multi-dimensional status data, including at least temperature data of the heat concentration area, high-frequency harmonic signals of the line current, and environmental parameters inside the distribution cabinet.

[0047] The data parsing unit is connected to the acquisition unit and is used to parse the multi-dimensional state data;

[0048] The edge computing unit is connected to the data parsing unit, and is used to run at least one AI analysis model on the parsed multi-dimensional state data, and output at least one device state analysis result corresponding to the multi-dimensional state data.

[0049] The decision unit is connected to the edge computing unit and is used to perform information fusion on the status analysis results of at least one of the devices based on preset rules to output fault information, including fault type and fault location.

[0050] The control unit is connected to the decision unit and is used to control the execution module in the power distribution cabinet to perform actions based on the fault information obtained by communication transmission.

[0051] The analysis unit is connected to the control unit and is used to calculate the fault suppression success rate within a preset period, determine the fault detection status based on the fault suppression success rate, adjust the transmission rate between the decision unit and the control unit based on the fault detection status, and adjust the startup cycle of the model incremental learning module based on the fault detection status after adjusting the transmission rate. The fault suppression success rate is the ratio of the number of times the potential fault is eliminated to the number of actions of the execution module. The model incremental learning module is a module in the edge computing unit that optimizes the AI ​​model using newly generated data.

[0052] Specifically, the multi-sensor array in the acquisition unit includes a cluster of temperature sensors that collect temperature data from multiple heat concentration zones, a high-frequency current transformer that collects high-frequency harmonic signals of line current, and an environmental sensor that collects environmental parameters within the cabinet.

[0053] Specifically, the data parsing unit parses multi-dimensional state data, including preprocessing and feature extraction of the data, including filtering the multi-dimensional state data to eliminate noise interference, cleaning the temperature data to remove outliers, and calculating the feature values ​​of different types of data in the multi-dimensional state data. This process is existing technology and will not be described in detail here.

[0054] Specifically, the AI ​​analysis models in the edge computing unit include a temperature fault model, a partial discharge identification model, and an environmental risk model. The temperature fault model outputs device status analysis results based on a combination of rule-based state judgment and trend prediction algorithms. The rule-based state judgment includes setting fixed alarm thresholds; calculating the relative temperature difference at the same location at different times within the same circuit; and setting alarm thresholds for the relative temperature difference to identify potential contact problems. The trend prediction algorithm uses linear regression analysis of historical temperature data to predict short-term temperature trends and determines whether to issue a warning based on the rule-based state judgment.

[0055] Specifically, the partial discharge identification model employs a deep learning approach, using a pre-trained one-dimensional convolutional neural network model to perform end-to-end analysis of the original waveforms of high-frequency current or ultrasonic signals. It automatically classifies the discharge type and assesses its severity level, comprehensively utilizes the high-frequency current signal to determine the discharge phase and intensity, and uses the time difference of the ultrasonic signal received by multiple ultrasonic sensors for time difference positioning to achieve precise positioning of the discharge power supply within the cabinet space, thereby outputting the equipment status analysis results. This process is existing technology and will not be elaborated further.

[0056] Specifically, the algorithm principle of the environmental risk model is based on logical judgment and threshold triggering. It compares environmental parameters with thresholds to output equipment status analysis results.

[0057] Specifically, the preset rules include setting threshold conditions for monitoring parameters and setting logical combinations for fault determination. The monitoring parameters include contact temperature, relative temperature difference, environmental parameters, partial discharge characteristic quantity and partial discharge intensity. The logical combination is defined as triggering the fault type and fault location corresponding to the combination when the actual data of one or more of the monitoring parameters meet their corresponding threshold conditions.

[0058] Please see Figure 2 The diagram shows a flowchart of the detection method for an intelligent power distribution cabinet with fault detection function according to an embodiment of the present invention.

[0059] S1, by acquiring multi-dimensional status data through a multi-sensor array installed in the distribution cabinet, including at least temperature data of the heat concentration area, high-frequency harmonic signals of line current and environmental parameters in the distribution cabinet;

[0060] S2, the multi-dimensional state data is parsed by the data parsing unit connected to the acquisition unit;

[0061] S3, run at least one AI analysis model on the parsed multi-dimensional state data through the edge computing unit connected to the data parsing unit, and output at least one device state analysis result corresponding to the multi-dimensional state data;

[0062] S4, the decision unit connected to the edge computing unit performs information fusion on the status analysis results of at least one device based on preset rules to output fault information, including fault type and fault location;

[0063] S5, the control unit connected to the decision unit controls the execution module in the power distribution cabinet to perform actions based on the fault information obtained through communication transmission;

[0064] S6, the analysis unit connected to the control unit calculates the fault suppression success rate within a preset period, determines the fault detection status based on the fault suppression success rate, and adjusts the transmission rate between the decision unit and the control unit based on the fault detection status, and adjusts the start-up period of the model incremental learning module based on the fault detection status after adjusting the transmission rate. The fault suppression success rate is the ratio of the number of times the potential fault is eliminated to the number of actions of the execution module, and the model incremental learning module is the module in the edge computing unit that optimizes the AI ​​model using newly generated data.

[0065] Please see Figure 3 The diagram shows a flowchart illustrating the steps of determining the fault suppression success rate based on a comparison with a preset value in an embodiment of the present invention. The analysis unit in this embodiment is further configured to calculate the difference between the preset value and the fault suppression success rate within a preset period when the fault detection state is determined to be unqualified; the analysis unit is further configured to adjust the transmission rate between the decision unit and the control unit based on the ratio of the difference to the preset difference when the difference is greater than the preset difference; wherein, when the fault suppression success rate is less than the preset value, the fault detection state is unqualified.

[0066] Specifically, taking the fault early warning of intelligent power distribution cabinet as an example, and based on the hardware performance limits of its multi-sensor array and the fault tolerance requirements of the actual operating environment inside the cabinet, as well as combining several historical data obtained through statistics and analysis during historical operation and maintenance, the corresponding preset or critical parameter values ​​are set.

[0067] Specifically, if the preset value L0 = 0.8, the comparison process between the fault suppression success rate L and the preset value L0 is as follows:

[0068] If the fault suppression success rate L is greater than or equal to the preset value L0, the fault detection status is deemed qualified.

[0069] If the fault suppression success rate L is less than the preset value L0, the fault detection status is deemed unqualified.

[0070] Specifically, when the fault detection status is determined to be unqualified, the difference between the preset value and the fault suppression success rate within the preset period is calculated. If the difference is greater than the preset difference, it indicates that there is a time lag in the fault information received by the control unit from the decision unit, causing the control unit to act too late. In this case, the transmission rate between the decision unit and the control unit needs to be adjusted based on the ratio of the difference to the preset difference. The preset ratio of the difference to the preset difference is P0 = 1.9. The comparison process based on the ratio P of the difference to the preset difference and the preset ratio P0 is as follows:

[0071] If the ratio P of the difference to the preset difference is less than or equal to the preset ratio P0, the transmission rate between the decision unit and the control unit will be adjusted to 1.76 times the original transmission rate.

[0072] If the ratio P of the difference to the preset difference is greater than the preset ratio P0, then the transmission rate between the decision unit and the control unit will be adjusted to 2.95 times the original transmission rate.

[0073] Specifically, the decision-making unit also includes a rule-establishing module for constructing preset rules by establishing different rule parameter sets according to different time periods. The fault detection status is redefined after adjusting the transmission rate between the decision-making unit and the control unit. If the fault detection status is unqualified, the fault suppression success rate at multiple historical moments is obtained. A historical period-fault suppression success rate curve is plotted, and the autocorrelation function is used to determine whether the curve has periodicity. The autocorrelation function determines periodicity by calculating the similarity between the signal and itself at different lag periods: if the curve shows a significant peak at a specific lag period, it indicates the existence of periodicity. This method is existing technology and will not be elaborated further.

[0074] Specifically, if the curve is periodic, it indicates that the user's electricity load has obvious daily or weekly patterns, such as high during the day and low at night, high on weekdays and low on weekends. However, the current preset rules are fixed and fail to adapt to this change. In other words, the preset rules do not match the fault analysis at the current stage and the preset rules need to be updated based on the current situation. Therefore, the rule establishment module is activated.

[0075] Specifically, the fault detection status is redefined. If the fault detection status is unqualified, the integral of the historical cycle-fault suppression success rate curve is calculated. If the integral is less than the preset integral, it indicates that the AI ​​analysis model has low accuracy and generates a large number of false alarms or missed alarms because it has not optimized the model based on the newly generated data. In this case, the start-up cycle of the model incremental learning module needs to be adjusted based on the ratio of the preset integral to the preset integral. The preset ratio of the preset integral to the preset integral is Q0=1.7. The comparison process between the preset integral to the preset integral ratio Q and the preset ratio Q0 is as follows:

[0076] If the ratio of the preset integral to the integral Q is less than or equal to the preset ratio Q0, the startup cycle of the model incremental learning module will be adjusted to 0.83 times the original startup cycle.

[0077] If the ratio Q of the preset integral to the original integral is greater than the preset ratio Q0, the startup cycle of the incremental learning module of the model will be adjusted to 0.69 times the original startup cycle.

[0078] Please see Figure 4 The diagram shown is a flowchart illustrating the steps for determining the fault detection status based on adjusting the startup cycle of the incremental learning module in an embodiment of the present invention.

[0079] Specifically, after re-determining the fault detection status after adjusting the startup cycle of the incremental learning module, if the fault detection status is unqualified, the startup cycle of the incremental learning module is adjusted at least once until the number of adjustments is less than the preset number and the fault detection status is qualified, or the number of adjustments is equal to the preset number, at which point the adjustment stops. After stopping the adjustment, if the fault detection status is unqualified, the variance of the fault suppression success rate at multiple historical moments is calculated. If the variance is less than the preset variance, it indicates that the system has missing fault data, the variance is lower than the preset value, and the number of sensors deployed in key parts is insufficient, resulting in the inability to eliminate outliers through data verification and missing data, thus making the acquired multi-dimensional status data inaccurate. In this case, it is necessary to adjust the number of different sensors in the multi-sensor array based on the difference between the variance and the preset variance. The preset difference between the two variances is R0=0.1. The comparison process based on the difference between the two variances R and the preset difference R0 is as follows:

[0080] If the difference R between the preset variance and the original variance is less than or equal to the preset difference R0, then the number of different sensors in the multi-sensor array will be adjusted to 1.5 times the original number, and the adjusted values ​​will be rounded up.

[0081] If the difference R between the preset variance and the original variance is greater than the preset difference R0, the number of different sensors in the multi-sensor array will be adjusted to 2.7 times the original number, and the adjusted values ​​will be rounded up.

[0082] Specifically, the more sensors there are, the greater the probability of individual sensor failures, such as data drift, complete failure, or communication interruption. To ensure the overall reliability of the data provided by the multi-sensor array, the "health status" of each sensor must be checked more frequently to promptly identify and remove faulty sensors and prevent their erroneous data from affecting the overall system's decision-making. Therefore, the sensor check cycle needs to be adjusted based on the difference between the preset number of sensors in the multi-sensor array and the preset difference T0, where the preset difference T0 = 2. The comparison process based on the difference T0 is as follows:

[0083] If the difference T between the preset quantity and the quantity is less than or equal to the preset difference T0, the sensor's verification cycle will be adjusted to 0.7 times the original verification cycle, where the verification cycle is in hours, and the adjusted values ​​will be rounded up.

[0084] If the difference T between the preset quantity and the actual quantity is greater than the preset difference T0, the sensor's verification cycle will be adjusted to 0.86 times the original verification cycle, where the verification cycle is in hours, and the adjusted values ​​will be rounded up.

[0085] 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.

[0086] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An intelligent power distribution cabinet with fault detection function, characterized in that, include: The acquisition unit uses a multi-sensor array installed in the distribution cabinet to collect multi-dimensional status data, including at least temperature data of the heat concentration area, high-frequency harmonic signals of line current and environmental parameters inside the distribution cabinet. A data parsing unit, which is connected to the acquisition unit, is used to parse the multi-dimensional state data; An edge computing unit, connected to the data parsing unit, is used to run at least one AI analysis model on the parsed multi-dimensional state data and output at least one device state analysis result corresponding to the multi-dimensional state data. A decision unit, connected to the edge computing unit, is used to perform information fusion on the status analysis results of at least one of the devices based on preset rules, so as to output fault information, including fault type and fault location. A control unit, which is connected to the decision unit, is used to control the execution module in the power distribution cabinet to perform actions based on the fault information obtained through communication transmission; An analysis unit, connected to the control unit, is used to calculate the fault suppression success rate within a preset period, determine the fault detection status based on the fault suppression success rate, adjust the transmission rate of fault information from the decision unit to the control unit based on the fault detection status, and adjust the startup cycle of the model incremental learning module based on the fault detection status after adjusting the transmission rate. The fault suppression success rate is the ratio of the number of times the potential fault is eliminated to the number of actions of the execution module. The model incremental learning module is a module in the edge computing unit that optimizes the AI ​​analysis model using newly generated data.

2. The intelligent power distribution cabinet with fault detection function according to claim 1, characterized in that, The analysis unit is also used to calculate the difference between a preset value and the fault suppression success rate within a preset period when it is determined that the fault detection status is unqualified. The analysis unit is also used to adjust the transmission rate of fault information from the decision unit to the control unit based on the ratio of the difference to the preset difference when the difference is greater than the preset difference. Wherein, if the success rate of fault suppression is less than the preset value, the fault detection status is unqualified.

3. The intelligent power distribution cabinet with fault detection function according to claim 2, characterized in that, The analysis unit is also used to increase the transmission rate of fault information from the decision unit to the control unit based on the ratio of the difference to a preset difference, and the increase in transmission rate is proportional to the ratio.

4. The intelligent power distribution cabinet with fault detection function according to claim 3, characterized in that, The decision-making unit also includes a rule-establishing module, which is used to establish different rule parameter sets according to different time periods in order to construct the preset rules; The analysis unit also includes obtaining the fault suppression success rate for multiple historical periods when the fault detection status is unqualified after adjusting the transmission rate. The analysis unit is also used to determine whether the plotted historical cycle-fault suppression success rate curve has periodicity by using the autocorrelation function. The decision-making unit is also used to activate the rule-establishing module when the curve has periodicity.

5. The intelligent power distribution cabinet with fault detection function according to claim 4, characterized in that, The analysis unit is also used to calculate the integral of the historical cycle-fault suppression success rate curve when the fault detection status is unqualified after the rule establishment module is started. The analysis unit is also used to adjust the startup cycle of the model incremental learning module based on the ratio of the preset integral to the integral when the integral is less than the preset integral.

6. The intelligent power distribution cabinet with fault detection function according to claim 5, characterized in that, The analysis unit is also used to reduce the startup cycle of the model incremental learning module based on the ratio of a preset integral to the integral, and the reduction in startup cycle is proportional to the ratio.

7. The intelligent power distribution cabinet with fault detection function according to claim 6, characterized in that, The analysis unit is also used to repeatedly adjust the startup cycle of the model incremental learning module at least once if the fault detection status is not qualified after adjusting the startup cycle of the model incremental learning module, until the number of adjustments is less than the preset number and the fault detection status is qualified or the number of adjustments is equal to the preset number, and then stop adjusting. The analysis unit is also used to calculate the variance of the fault suppression success rate over multiple historical periods when the fault detection status is unqualified after the adjustment is stopped. The analysis unit is also used to adjust the number of different sensors in the multi-sensor array based on the difference between the preset variance and the variance when the variance is less than the preset variance.

8. The intelligent power distribution cabinet with fault detection function according to claim 7, characterized in that, The analysis unit is also used to increase the number of different sensors in the multi-sensor array based on the difference between the preset variance and the variance, and the increase in the number of different sensors in the multi-sensor array is proportional to the difference.

9. The intelligent power distribution cabinet with fault detection function according to claim 8, characterized in that, The analysis unit is also used to reduce the sensor verification cycle based on the difference between the preset number of sensors in the multi-sensor array and the actual number of sensors, and the reduction in the verification cycle is inversely proportional to the difference.

10. A detection method applied to an intelligent power distribution cabinet with fault detection function as described in any one of claims 1-9, characterized in that, include: The acquisition unit uses a multi-sensor array installed in the distribution cabinet to collect multi-dimensional status data, including at least temperature data of the heat concentration area, high-frequency harmonic signals of line current and environmental parameters inside the distribution cabinet. The multi-dimensional state data is analyzed by a data parsing unit connected to the acquisition unit; The edge computing unit connected to the data parsing unit runs at least one AI analysis model on the parsed multi-dimensional state data and outputs at least one device state analysis result corresponding to the multi-dimensional state data. The decision unit connected to the edge computing unit performs information fusion on the status analysis results of at least one of the devices based on preset rules to output fault information, including fault type and fault location. The control unit connected to the decision unit controls the execution module in the power distribution cabinet to perform actions based on the fault information obtained through communication transmission. The analysis unit connected to the control unit calculates the fault suppression success rate within a preset period, determines the fault detection status based on the fault suppression success rate, adjusts the transmission rate of fault information from the decision unit to the control unit based on the fault detection status, and adjusts the startup cycle of the model incremental learning module based on the fault detection status after adjusting the transmission rate. The fault suppression success rate is the ratio of the number of times the potential fault is eliminated to the number of actions of the execution module. The model incremental learning module is a module in the edge computing unit that optimizes the AI ​​analysis model using newly generated data.

Images