Modular intelligent distribution box and intelligent control method thereof

Abstract

This invention relates to the field of distribution box technology, and proposes a modular intelligent distribution box and its intelligent control method, including: arranging monitoring points for each module of the distribution box and collecting temperature data; obtaining several initial time periods for each monitoring point; obtaining the probability of sudden changes in the operating state of each monitoring point during each initial time period, and thus obtaining several target time periods for each monitoring point; obtaining several local abnormal regions at each moment; obtaining several abnormal regions and their abnormal time periods; obtaining the abnormal recovery index of the abnormal regions; obtaining the true abnormal probability of each abnormal region; obtaining the load limiting index of each local abnormal region at the current moment, and then limiting the load of the corresponding modules of each monitoring point in the local abnormal region. This invention aims to solve the problem that the arbitrary plugging and unplugging of modules in the distribution box disrupts heat distribution, leading to misjudgments and affecting control.

Description

Technical Field

[0001] This invention relates to the field of distribution box technology, and specifically to a modular intelligent distribution box and its intelligent control method. Background Technology

[0002] With the increasing demands for power supply flexibility and operational efficiency in industrial development, modular intelligent distribution boxes are widely used in practical industrial scenarios due to their advantages such as supporting hot-swapping of distribution units, easy expansion, and flexible reconfiguration. To ensure power safety, existing modular intelligent distribution boxes are typically equipped with intelligent environmental and temperature control systems. Currently, temperature control and safety monitoring solutions usually involve placing temperature sensors inside the distribution box or near the module slots to monitor the distribution box's operating status in real time. Based on historical operating references, thresholds are set. When the monitored temperature exceeds the threshold, the intelligent system triggers an early warning or control mechanism, such as dissipating heat from the abnormally heated area or cutting off the power to the abnormal module, thereby achieving intelligent control of the modular distribution box.

[0003] However, in the actual operation of modular intelligent distribution boxes, the physical topology and specific thermal distribution inside the box are dynamically fluctuating. Due to the characteristic that its modules can be plugged in and out at will, the moment of plugging and unplugging may disrupt the original heat distribution state inside the box, causing local heat to accumulate passively. This may cause existing methods to misjudge the local accumulation state with the actual abnormal temperature rise (such as the underlying electrical hidden danger of continuous heating due to excessive contact resistance when the connector contacts deteriorate), thus leading to the output of incorrect control commands. Summary of the Invention

[0004] This invention provides a modular intelligent distribution box and its intelligent control method to solve the problem that the heat distribution is disrupted due to the arbitrary plugging and unplugging of modules in existing distribution boxes, leading to misjudgments and affecting control. The specific technical solution adopted is as follows: This invention proposes an intelligent control method for a modular intelligent distribution box, which includes the following steps: Monitoring points were set up for each module of the distribution box and temperature was collected. The topology of the monitoring points was obtained based on the distribution of the modules and monitoring points. Based on the time-series changes in the monitored temperature at the same monitoring point, several initial time periods for each monitoring point are obtained; the differences in the monitoring temperature change trends between adjacent initial time periods are analyzed, and combined with the monitoring temperature performance and fluctuations within the initial time periods, the probability of sudden changes in the operating status of each monitoring point in each initial time period is obtained, thereby obtaining several target time periods for each monitoring point; based on the distribution of target time periods for each monitoring point, several local abnormal areas at each moment are obtained. Based on the distribution of local abnormal regions over continuous time, several abnormal regions and their abnormal time periods are obtained; the changes in the probability of sudden changes in the operating status of each monitoring point in the abnormal region during the target time period are analyzed, and the abnormal recovery index of the abnormal region is obtained by combining the changes in the area and position of the abnormal region; considering the similarity between the direction of position change of different abnormal regions at the same time, the true abnormal probability of each abnormal region is obtained by combining the abnormal recovery index. Based on the number of monitoring points in the local anomaly area at the current moment and the probability of sudden changes in their operating status, combined with the historical actual anomaly probability performance of the local anomaly area, the load limitation index of each local anomaly area at the current moment is obtained, and then the load of the corresponding module of each monitoring point in the local anomaly area is limited.

[0005] Optionally, the specific method for obtaining several initial time periods for each monitoring point includes: For any monitoring point, obtain the monitoring temperature of that monitoring point at each time point within the past week except for the most recent day, and construct the historical time series curve of that monitoring point; Several maximum points are obtained from the historical time series curve, and the historical time series curve is divided into several window curves by the maximum points. The number of time points in each window curve is obtained as the time series length of each window curve. The average of the time series lengths of the other window curves except the first window curve and the last window curve is used as the sliding window length of the monitoring point. The current time-series curve formed by the monitored temperatures at each moment of the most recent day at the monitoring point is divided into several non-overlapping segments using the sliding window length. The time periods corresponding to each curve segment are used as several initial time periods for the monitoring point.

[0006] Optionally, the specific method for obtaining the probability of sudden changes in the operating state of each monitoring point in each initial time period includes: For any monitoring point and any initial time period, the slope of each time in the initial time period is obtained based on the temperature difference between adjacent times. The average slope of all times in the initial time period is obtained as the average slope of the initial time period. The ratio of the average slope of the initial time period to the average slope of the adjacent previous initial time period is used as the first ratio of the initial time period. The ratio of the average monitored temperature at all times during the initial period at this monitoring point to the average monitored temperature in the historical time series curve of this monitoring point is taken as the second ratio for the initial period. Based on the first ratio, the second ratio, and the variance of the monitored temperature at all times in the initial period, the probability of a sudden change in the operating state during the initial period is obtained; the probability of a sudden change in the operating state is positively correlated with the first ratio, the second ratio, and the variance of the monitored temperature.

[0007] Optionally, the specific method for obtaining several target time periods for each monitoring point includes: A preset anomaly threshold is set, and the initial time period in which the probability of a sudden change in the operating state of any monitoring point is greater than the anomaly threshold is taken as the initial target time period of that monitoring point. Among the several initial target time periods of the monitoring point, consecutive initial target time periods are merged, and all time periods that are not included in the merging and those after merging are taken as several target time periods of the monitoring point.

[0008] Optionally, the specific method for obtaining several local anomaly regions at each time point includes: For any monitoring point, at any time within the target time period, the monitoring point at that time is recorded as the anomaly detection point at that time; all anomaly monitoring points at that time are obtained, and based on the monitoring point topology, several local anomaly regions at that time are constructed from all the anomaly monitoring points at that time.

[0009] Optionally, the specific methods for obtaining several abnormal regions and their abnormal time periods include: For any two adjacent moments in the most recent day, the two local anomaly regions with the largest overlapping area in the adjacent moments are taken as the anomaly regions of the two adjacent moments. The local anomaly regions in consecutive moments are matched according to the overlapping area to obtain several anomaly regions in consecutive adjacent moments. For any abnormal region and its corresponding consecutive adjacent time period, the consecutive adjacent time period is taken as the abnormal time period of the abnormal region.

[0010] Optionally, the specific method for obtaining the anomaly recovery index of the anomaly region includes: For any monitoring point in any abnormal region, obtain several target time periods corresponding to the abnormal time period of the monitoring point in the abnormal region, as several abnormal target time periods of the monitoring point in the abnormal region. Obtain the ratio of the sum of the time lengths of all the abnormal target time periods to the time length of the abnormal time period and use it as the third ratio. The ratio of the probability of the sudden change in the operating state of the monitoring point in the first abnormal target time period to the probability of the sudden change in the operating state in the last abnormal target time period in all the abnormal target time periods, and the result obtained by comparing it with the third ratio, is used as the abnormal fall-off index of the monitoring point. Obtain the center point of the abnormal region at the first moment and the center point at the last moment during the abnormal period, and denote them as the starting center point and the ending center point, respectively. Use the vector obtained by pointing the starting center point to the ending center point as the abnormal transmission vector of the abnormal region. Obtain the ratio of the area of ​​the local abnormal region corresponding to the next moment to the area of ​​the local abnormal region corresponding to the previous moment in any adjacent moment of the abnormal region. Based on the sum of the ratios obtained at all adjacent time points, the mean of the anomaly fall-off index of all monitoring points in the anomaly region, and the magnitude of the anomaly propagation vector, the anomaly recovery index of the anomaly region is obtained; the anomaly recovery index is positively correlated with the sum of the obtained ratios and the mean of the anomaly fall-off index, and negatively correlated with the magnitude.

[0011] Optionally, the specific method for obtaining the true anomaly probability of each anomaly region includes: For any local abnormal region at any time during an abnormal period of any abnormal region, the abnormal transmission vector of the abnormal region is assigned to the local abnormal region at that time. The cosine similarity between the abnormal transmission vector of the local abnormal region at that time and the abnormal transmission vectors of all other local abnormal regions is obtained. The reciprocal of the distance between the center points of two local abnormal regions is used as the weight to calculate the weighted average of all cosine similarities. The result is used as the reference factor for the position change of the abnormal region at that time during the abnormal period. Based on the anomaly recovery index of the abnormal region and the mean of the location change reference factor at all times during the abnormal period, the true anomaly probability of the abnormal region is obtained. The true anomaly probability is negatively correlated with both the anomaly recovery index and the mean of the location change reference factor. The true anomaly probability of the abnormal region is assigned to the corresponding local abnormal region at each time step in the abnormal region, thereby obtaining the true anomaly probability of each local abnormal region at each time step.

[0012] Optionally, the specific method for obtaining the load limitation index of each local abnormal region at the current time includes: For any local anomaly region at the current moment, obtain the ratio of the actual anomaly probability of the local anomaly region at the current moment to the average of the actual anomaly probabilities of the local anomaly region at all historical moments; based on the ratio, the number of monitoring points in the local anomaly region at the current moment, and the average of the probability of sudden changes in the operating status of each monitoring point at the current moment, obtain the load limitation index of the local anomaly region at the current moment.

[0013] The present invention also proposes a modular intelligent distribution box, which includes a distribution box, a monitoring unit, a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the above method to intelligently control the distribution box.

[0014] The beneficial effects of this invention are as follows: This invention filters the time periods and spatial distribution of abnormal temperature fluctuations within the modular intelligent distribution box based on the actual monitored temperatures at each monitoring point. Based on the spatiotemporal recovery characteristics of the abnormal range under the actual working conditions within the corresponding box and the stability of the abnormal temperature diffusion trend, it distinguishes between temporary abnormal heat fluctuations and actual defect temperature accumulation. Furthermore, it determines the corresponding control index based on the degree of abnormal accumulation, thereby eliminating the impact of local temperature mutations caused by hot-plugging on the overall intelligent temperature control of the modular intelligent distribution box. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic flowchart of a modular intelligent distribution box and its intelligent control method provided in one embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1 The diagram illustrates a modular intelligent distribution box and its intelligent control method according to an embodiment of the present invention. The method includes the following steps: Step S001: Set up monitoring points for each module of the distribution box and collect monitoring temperatures, and obtain the monitoring point topology based on the distribution of modules and monitoring points.

[0019] The purpose of this embodiment is to decouple the abnormal temperature rise state by extracting the thermal recovery capacity and spatiotemporal diffusion degree of the abnormal heat accumulation area under different operating conditions, thereby distinguishing between temporary abnormal heat state fluctuations and real defect temperature rise accumulation, and determining the actual control index of the modular distribution box based on the accumulation degree of its real abnormal evolution.

[0020] It should be noted that, based on the spatiotemporal fluctuation characteristics of the abnormal temperature region, the abnormal temperature caused by the operating condition disturbance is distinguished from the actual temperature abnormality, so as to determine the actual degree of control of different modules according to the specific attribution.

[0021] Specifically, monitoring points are set up for the modular distribution box: since most defects occur due to connector oxidation, monitoring points are set up at the "interlocking points of moving and stationary contacts (inlet and outlet terminals)" of the physical slots of each module; specifically, monitoring points are set up at the physical isolation plates between adjacent slots and at the inlet and outlet of the main convection ducts; temperature sensors are installed at each monitoring point to monitor the temperature data at the corresponding location in real time, thereby obtaining the continuous monitoring temperature of each monitoring point; at the same time, miniature Hall current sensors are integrated at the interfaces of each module to capture the real-time load of the module; the collected monitoring temperature and real-time load of the module are transmitted to the data processing unit through the bus to clean the data and ensure that the timestamps between the data are aligned; each module is matched with its nearest sensor (and the coordinates of each sensor are recorded), and its topology is obtained according to the actual physical location of each module, thereby obtaining the topology of the monitoring points.

[0022] It should be noted that in modular intelligent distribution boxes, local abnormal fluctuations may occur due to factors such as dynamic plugging and unplugging of modules, load fluctuation interference from adjacent modules, and contact aging. Existing methods are difficult to decouple and control the above-mentioned abnormal coupling, which may lead to misjudgment of thermal anomalies caused by temporary operating condition interference and real faults. Therefore, based on the recovery degree of local heat accumulation abnormal disturbance and spatiotemporal correlation, temporary operating condition disturbance anomalies are distinguished, and corresponding control is carried out on different abnormal modules based on the distinction results.

[0023] Step S002: Based on the time-series changes in the monitored temperature at the same monitoring point, several initial time periods are obtained for each monitoring point; the differences in the monitoring temperature change trends between adjacent initial time periods are analyzed, and the monitoring temperature performance and fluctuations within the initial time periods are combined to obtain the probability of sudden changes in the operating status of each monitoring point in each initial time period, thereby obtaining several target time periods for each monitoring point; based on the distribution of target time periods for each monitoring point, several local abnormal areas at each moment are obtained.

[0024] It should be noted that in the actual operation of modular intelligent distribution boxes, the specific operating status of different modules and their actual power load have a certain degree of randomness. This causes the specific heat accumulation and heat diffusion distribution at a single monitoring location inside the box to be continuously affected by the operating status of neighboring modules, resulting in dynamic fluctuations. When the status of the modules in the vicinity is abnormal, the temperature accumulation at the corresponding monitoring location will fluctuate abnormally. Therefore, it is necessary to determine the target time period for different monitoring points based on the time-series temperature fluctuation of the data obtained from each monitoring point.

[0025] Preferably, in one embodiment of the present invention, several initial time periods for each monitoring point are obtained based on the temporal changes in the monitored temperature at the same monitoring point, including the following specific method: It should be further noted that since the neighboring modules of a single monitoring point are of similar types, their load fluctuation trends also have certain changing trends. Therefore, the sliding window size is determined based on the temperature fluctuation trend captured by the monitoring point in the past week (the measured values ​​of local monitoring points are affected by the ambient temperature, and the load changes of neighboring modules usually have seasonal characteristics, so short-term monitoring data is used as historical data for reference). This allows for the initial division of time periods.

[0026] Specifically, for any monitoring point, the monitored temperature at each time point within the past week, excluding the most recent day, is obtained, forming a historical time-series curve for that monitoring point. Several maximum points are obtained from the historical time-series curve, and the curve is divided into several window curves based on these maximum points. The number of times in each window curve is used as the time-series length of each window curve. The average time-series length of all window curves except the first and last window curves is used as the sliding window length for that monitoring point. The current time-series curve formed by the monitored temperatures at each time point on the most recent day is divided into several non-overlapping segments using the sliding window length. The time periods corresponding to each segment are used as several initial time periods for that monitoring point. It is worth noting that if the number of remaining times is less than the sliding window length, the remaining times are merged with the last obtained initial time period and used again as initial time periods for subsequent analysis.

[0027] It should be further noted that the temperature change at a single monitoring point is usually relatively gradual over a short period of time. If the temperature monitoring value shows an abnormal increase within a certain period (i.e., the temperature increase is large in adjacent initial periods and is higher than the temperature in historical monitoring periods) and is accompanied by unstable fluctuations, there is a high probability that there is a sudden change in the operating status at the current monitoring point.

[0028] Preferably, in one embodiment of the present invention, the differences in the monitoring temperature change trends of adjacent initial time periods of the monitoring points are analyzed, and the monitoring temperature performance and fluctuations within the initial time period are combined to obtain the probability of sudden changes in the operating state of each monitoring point in each initial time period, thereby obtaining several target time periods for each monitoring point. The specific method includes: For any monitoring point and any initial time period, the slope of each moment in the initial time period is obtained based on the temperature difference between adjacent moments. The average slope of all moments in the initial time period is obtained as the average slope of the initial time period. The ratio of the average slope of the initial time period to the average slope of the adjacent previous initial time period is used as the first ratio of the initial time period. The ratio of the average temperature of all moments in the initial time period of the monitoring point to the average temperature of all moments in the historical time series curve of the monitoring point is used as the second ratio of the initial time period. Based on the first ratio, the second ratio, and the variance of the monitoring temperature at all moments in the initial time period, the probability of a sudden change in the operating state of the initial time period is obtained. The probability of a sudden change in the operating state is positively correlated with the first ratio, the second ratio, and the variance of the monitoring temperature.

[0029] As an example, this embodiment uses the product of the first ratio, the second ratio, and the normalized result of the monitoring temperature variance at all times in the initial period as the probability of sudden change in the operating state in the initial period. The normalization adopts the maximum and minimum value normalization, and the normalization object is the product obtained for all time periods.

[0030] Furthermore, an abnormal threshold is preset, and in this embodiment, the abnormal threshold is described as 0.5; the initial time period in which the probability of a sudden change in the operating state of any monitoring point is greater than the abnormal threshold is taken as the initial target time period of the monitoring point; consecutive initial target time periods in the initial target time periods of the monitoring point are merged, and all time periods that are not merged and those after merging are taken as the target time periods of the monitoring point.

[0031] It should be noted that the first ratio reflects the difference in temperature change amplitude between adjacent initial time periods, and the second ratio reflects the monitoring temperature amplitude performance of the corresponding initial time period. Combined with the monitoring temperature fluctuation performance within the initial time period, the probability of sudden change in motion state is quantified. Then, the initial time period is screened by threshold and consecutive initial time periods are merged to obtain the target time period.

[0032] It should be further explained that after the monitoring points and the corresponding target time periods that may be abnormal are extracted, the monitoring points at individual moments are further analyzed as a whole, that is, the local abnormal areas that may occur at a single moment are obtained.

[0033] Preferably, in one embodiment of the present invention, based on the target time period distribution of each monitoring point, several local abnormal regions at each time moment are obtained, including the following specific method: For any monitoring point, at any time within the target time period, the monitoring point at that time is recorded as the anomaly detection point at that time; all anomaly monitoring points at that time are obtained, and based on the monitoring point topology, several local anomaly regions at that time are constructed from all the anomaly monitoring points at that time.

[0034] Thus, several local anomaly regions at each time point were obtained.

[0035] Step S003: Based on the distribution of local abnormal regions at continuous time intervals, obtain several abnormal regions and their abnormal time periods; analyze the change in the probability of sudden changes in the operating status of each monitoring point in the abnormal region during the target time period, and obtain the abnormal recovery index of the abnormal region by combining the area and position changes of the abnormal region; consider the similarity between the position change directions of different abnormal regions at the same time, and obtain the true abnormal probability of each local abnormal region at each time interval by combining the abnormal recovery index.

[0036] It should be noted that during the actual operation of the modular distribution box, local heat distribution is affected by dynamic factors, which can cause abnormal fluctuations at the corresponding monitoring points. Specifically, the load of local modules in the distribution box fluctuates randomly with actual power demand. Changes in the load status of adjacent modules and normal operating conditions such as module insertion and removal may alter the local heat distribution. Abnormal increases in contact resistance due to oxidation or mechanical fatigue of the actual contact points of the module insertion may also cause abnormal local heat distribution. Therefore, it is necessary to distinguish between real abnormalities and normal operating condition fluctuations by considering the spatiotemporal fluctuation characteristics of abnormal temperature rise.

[0037] It should be further noted that changes in the distribution of hot airflow caused by disturbances under normal operating conditions may disrupt the local temperature steady state, potentially leading to a brief but subsidable abnormal temperature rise at local monitoring points. Furthermore, the direction of heat propagation is limited by the direction of heat transfer when the hot airflow path inside the chamber changes. Therefore, the abnormal area is relatively concentrated, appears along the airflow direction, and disappears quickly. In contrast, the abnormal point of a real fault experiences a stable abnormal temperature rise, which usually causes the temperature to accumulate continuously in the vicinity without significant drop, and gradually spreads to the surrounding areas, resulting in the abnormal range expanding over time. Moreover, the direction of propagation is not dominated by the airflow and exhibits a central divergent characteristic.

[0038] Preferably, in one embodiment of the present invention, several abnormal regions and their abnormal time periods are obtained based on the distribution of local abnormal regions at continuous time intervals, including the following specific method: For any two adjacent moments in the most recent day, the two local abnormal regions with the largest overlapping area in the adjacent moments are taken as the abnormal regions of the two adjacent moments. Local abnormal regions in consecutive moments are matched according to the overlapping area to obtain several abnormal regions of consecutive adjacent moments. For any abnormal region and its corresponding consecutive adjacent moments, the consecutive adjacent moments are taken as the abnormal time period of the abnormal region.

[0039] Preferably, in one embodiment of the present invention, the method of analyzing the change in the probability of sudden changes in the operating status of each monitoring point in the abnormal area during the target time period, and combining the changes in the area and location of the abnormal area, to obtain the abnormal recovery index of the abnormal area includes: For any monitoring point in any abnormal region, obtain several target time periods (the target time periods may not be continuous) corresponding to the monitoring point in the abnormal time period of the abnormal region, as several abnormal target time periods of the monitoring point in the abnormal region. Obtain the ratio of the sum of the time lengths of all the abnormal target time periods to the time length of the abnormal time period and use it as the third ratio. The ratio of the probability of the sudden change in the operating state of the monitoring point in the first abnormal target time period to the probability of the sudden change in the operating state in the last abnormal target time period in all the abnormal target time periods, and the result obtained by comparing it with the third ratio, is used as the abnormal decline index of the monitoring point.

[0040] It should be noted that the anomaly reduction index of a single monitoring point is as follows: if, within its corresponding target time period, the probability of a sudden change in its operating state decreases significantly as the sliding window moves, and the duration of its corresponding target time period and the abnormal time period of its local abnormal region are relatively small, then the anomaly reduction index of that monitoring point is larger: the larger the ratio of the probability of a sudden change in operating state between the first and last abnormal target time periods, the larger the reduction amplitude of the corresponding monitoring point; and the smaller the proportion of all abnormal target time periods of the monitoring point in the abnormal time period, the larger the corresponding anomaly reduction index.

[0041] It should be further noted that if the anomaly regression index of each monitoring point in a certain local anomaly area is large, and the area of ​​the area gradually decreases over time (i.e., the number of monitoring points in the area gradually decreases), and the anomaly transmission distance in the area is small, then the anomaly recovery index of the area is large.

[0042] Furthermore, the center points of the abnormal region at the first and last moments of the abnormal period are obtained and denoted as the starting center point and the ending center point, respectively. The vector obtained by pointing the starting center point to the ending center point is used as the abnormal transmission vector of the abnormal region. The ratio of the area of ​​the local abnormal region corresponding to the later moment to the area of ​​the local abnormal region corresponding to the earlier moment is obtained in any adjacent moment of the abnormal region. Based on the sum of the ratios obtained at all adjacent moments, the mean of the abnormal fall-off index of all monitoring points in the abnormal region, and the magnitude of the abnormal transmission vector, the abnormal recovery index of the abnormal region is obtained. The abnormal recovery index is positively correlated with the sum of the obtained ratios and the mean of the abnormal fall-off index, and negatively correlated with the magnitude.

[0043] As an example, in this embodiment, the product of the sum of the obtained ratios and the mean of the abnormal fall-off index is divided by the magnitude of the abnormal transmission vector, and the result is used as the abnormal recovery index of the abnormal region.

[0044] It should be further noted that if, during the same period, a region has a large abnormal recovery index and a high degree of similarity in the direction of abnormal spread with its neighboring local abnormal regions, then the probability that the abnormality in that region is a real defect abnormality as of the current moment is smaller.

[0045] Preferably, in one embodiment of the present invention, considering the similarity between the direction of positional change of different abnormal regions at the same time, and combining the abnormality recovery index to obtain the true probability of each local abnormal region at each time, the specific method includes: For any local abnormal region at any time during an abnormal period of any abnormal region, the abnormal transmission vector of the abnormal region is assigned to the local abnormal region at that time. The cosine similarity between the abnormal transmission vector of the local abnormal region at that time and the abnormal transmission vectors of all other local abnormal regions is obtained. The reciprocal of the distance between the center points of two local abnormal regions is used as the weight to calculate the weighted average of all cosine similarities. The result is used as the reference factor for the positional change of the abnormal region at that time during the abnormal period.

[0046] Furthermore, based on the anomaly recovery index of the abnormal region and the mean of the location change reference factor at all times during the abnormal period, the true anomaly probability of the abnormal region is obtained. The true anomaly probability is negatively correlated with both the anomaly recovery index and the mean of the location change reference factor.

[0047] As an example, this embodiment uses the inversely proportional normalized result of the product of the abnormal recovery index of the abnormal region and the mean of the location change reference factor at all times during the abnormal period as the true abnormality probability of the abnormal region. In this embodiment, the inverse proportional function is processed using the reciprocal, and the normalization adopts the maximum-minimum value normalization method. The normalization object is the inverse proportional value obtained for all abnormal regions.

[0048] Furthermore, the true anomaly probability of any anomalous region is assigned to the corresponding local anomalous region at each time step within that anomalous region, thereby obtaining the true anomaly probability of each local anomalous region at each time step.

[0049] Thus, the true probability of anomalies in each local anomaly region at each time point is obtained.

[0050] Step S004: Based on the number of monitoring points in the local abnormal area at the current moment and the probability of sudden changes in their operating status, combined with the historical true abnormality probability performance of the local abnormal area, the load limitation index of each local abnormal area at the current moment is obtained, and then the load of the corresponding module of each monitoring point in the local abnormal area is limited to realize the intelligent control of the modular distribution box.

[0051] It should be noted that when the local temperature anomaly of the modular distribution box is mainly caused by temporary disturbances such as obstruction of the air duct or fluctuations in normal operating conditions such as large loads nearby, physical environmental heat dissipation should be performed. If the corresponding anomaly is caused by an increase in actual contact resistance, the corresponding abnormal module is very likely to cause a fire under high current load, and the maximum allowable load of the abnormal module in that area needs to be limited and controlled.

[0052] Specifically, for any local anomaly region at the current moment, the ratio of the actual anomaly probability of the local anomaly region at the current moment to the average of the actual anomaly probabilities of the local anomaly region at all historical moments is obtained; based on the ratio, the number of monitoring points in the local anomaly region at the current moment, and the average of the probability of sudden changes in the operating status of each monitoring point at the current moment, the load limitation index of the local anomaly region at the current moment is obtained.

[0053] As an example, the product of the ratio, the number of monitoring points in the local anomaly region at the current time, and the mean of the probability of sudden changes in the operating status of each monitoring point at the current time, and the result of normalization, is used as the load limitation index of the corresponding local anomaly region at the current time. The normalization adopts the maximum-minimum value normalization method, and the normalization object is the product of the three corresponding values ​​obtained for all local anomaly regions at all times.

[0054] It should be noted that the true probability of anomalies at a historical moment is obtained based on the corresponding anomaly region. The anomaly region to which the local anomaly region belongs directly participates in the calculation. At other times, the method of matching the local anomaly regions with the largest overlapping area during the anomaly region acquisition process is used to match the anomaly regions of different anomaly periods, and then the true probability of obtaining the corresponding anomaly for the local anomaly region in other anomaly periods is calculated.

[0055] It should be noted that if the local anomaly area is smaller at the current moment, the less likely there is a sudden change in the operating status of each monitoring point, and the less likely it is to be a real defect anomaly and the more stable it is, the less the modules within its range need to be subject to load limiting.

[0056] Furthermore, a first control threshold and a second control threshold are preset. In this embodiment, the first control threshold is described as 0.5, and the second control threshold is described as 0.8. If the load limitation index of any local abnormal area at the current time is less than the first control threshold, the heat dissipation equipment of several modules corresponding to the local abnormal area is activated. If the load limitation index is greater than or equal to the second control threshold, the system directly outputs the warning information of several modules corresponding to the local abnormal area and triggers the circuit breaker of the corresponding module. If the load limitation index is greater than or equal to the first control threshold and less than the second control threshold, based on the rated load of the module corresponding to the local abnormal area, the load limitation index, the probability of a sudden change in the operating state of the nearest monitoring point of any module at the current time, and the ratio of the duration of the module in the local abnormal area to the total duration of all abnormal target time periods in the local abnormal area are introduced. The product of the load limitation index, the corresponding probability of a sudden change in the operating state, and the ratio is normalized by the sigmoid function, and the result is used as the adjustment parameter of the rated load for adjustment and limitation. The single adjustment amplitude cannot exceed the physical equipment limit. In this way, the load limitation index of the local abnormal area at the current time is used to intelligently control each module in the intelligent distribution box.

[0057] It should be noted that, in order to avoid the fraction being meaningless due to the denominator being 0 during the ratio and fraction calculation process in this embodiment, a hyperparameter is added to both the denominator and the numerator before the calculation. The hyperparameter in this embodiment is described as 0.1.

[0058] This concludes the embodiment.

[0059] Another embodiment of the present invention proposes a modular intelligent distribution box, which includes a distribution box, a monitoring unit, a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the above method to perform intelligent control of the distribution box.

[0060] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent control of a modular intelligent distribution box, characterized in that, The method includes the following steps: Monitoring points were set up for each module of the distribution box and temperature was collected. The topology of the monitoring points was obtained based on the distribution of the modules and monitoring points. Based on the time-series changes in the monitored temperature at the same monitoring point, several initial time periods for each monitoring point are obtained; the differences in the monitoring temperature change trends between adjacent initial time periods are analyzed, and combined with the monitoring temperature performance and fluctuations within the initial time periods, the probability of sudden changes in the operating status of each monitoring point in each initial time period is obtained, thereby obtaining several target time periods for each monitoring point; based on the distribution of target time periods for each monitoring point, several local abnormal areas at each moment are obtained. Based on the distribution of local abnormal regions over continuous time, several abnormal regions and their abnormal time periods are obtained; the changes in the probability of sudden changes in the operating status of each monitoring point in the abnormal region during the target time period are analyzed, and the abnormal recovery index of the abnormal region is obtained by combining the changes in the area and position of the abnormal region; considering the similarity between the direction of position change of different abnormal regions at the same time, the true abnormal probability of each abnormal region is obtained by combining the abnormal recovery index. Based on the number of monitoring points in the local anomaly area at the current moment and the probability of sudden changes in their operating status, combined with the historical actual anomaly probability performance of the local anomaly area, the load limitation index of each local anomaly area at the current moment is obtained, and then the load of the corresponding module of each monitoring point in the local anomaly area is limited.

2. The intelligent control method for a modular intelligent distribution box according to claim 1, characterized in that, The specific method for obtaining several initial time periods for each monitoring point is as follows: For any monitoring point, obtain the monitoring temperature of that monitoring point at each time point within the past week except for the most recent day, and construct the historical time series curve of that monitoring point; Several maximum points are obtained from the historical time series curve, and the historical time series curve is divided into several window curves by the maximum points. The number of time points in each window curve is obtained as the time series length of each window curve. The average of the time series lengths of the other window curves except the first window curve and the last window curve is used as the sliding window length of the monitoring point. The current time-series curve formed by the monitored temperatures at each moment of the most recent day at the monitoring point is divided into several non-overlapping segments using the sliding window length. The time periods corresponding to each curve segment are used as several initial time periods for the monitoring point.

3. The intelligent control method for a modular intelligent distribution box according to claim 2, characterized in that, The specific methods for obtaining the probability of sudden changes in the operating state of each monitoring point at each initial time period are as follows: For any monitoring point and any initial time period, the slope of each time in the initial time period is obtained based on the temperature difference between adjacent times. The average slope of all times in the initial time period is obtained as the average slope of the initial time period. The ratio of the average slope of the initial time period to the average slope of the adjacent previous initial time period is used as the first ratio of the initial time period. The ratio of the average monitored temperature at all times during the initial period at this monitoring point to the average monitored temperature in the historical time series curve of this monitoring point is taken as the second ratio for the initial period. Based on the first ratio, the second ratio, and the variance of the monitored temperature at all times in the initial period, the probability of a sudden change in the operating state during the initial period is obtained; the probability of a sudden change in the operating state is positively correlated with the first ratio, the second ratio, and the variance of the monitored temperature.

4. The intelligent control method for a modular intelligent distribution box according to claim 1, characterized in that, The specific methods for obtaining several target time periods for each monitoring point are as follows: A preset anomaly threshold is set, and the initial time period in which the probability of a sudden change in the operating state of any monitoring point is greater than the anomaly threshold is taken as the initial target time period of that monitoring point. Among the several initial target time periods of the monitoring point, consecutive initial target time periods are merged, and all time periods that are not included in the merging and those after merging are taken as several target time periods of the monitoring point.

5. The intelligent control method for a modular intelligent distribution box according to claim 1, characterized in that, The specific method for obtaining several local anomaly regions at each time point is as follows: For any monitoring point, at any time within the target time period, the monitoring point at that time is recorded as the anomaly detection point at that time; all anomaly monitoring points at that time are obtained, and based on the monitoring point topology, several local anomaly regions at that time are constructed from all the anomaly monitoring points at that time.

6. The intelligent control method for a modular intelligent distribution box according to claim 1, characterized in that, The specific methods for obtaining several abnormal regions and their abnormal time periods are as follows: For any two adjacent moments in the most recent day, the two local anomaly regions with the largest overlapping area in the adjacent moments are taken as the anomaly regions of the two adjacent moments. The local anomaly regions in consecutive moments are matched according to the overlapping area to obtain several anomaly regions in consecutive adjacent moments. For any abnormal region and its corresponding consecutive adjacent time period, the consecutive adjacent time period is taken as the abnormal time period of the abnormal region.

7. The intelligent control method for a modular intelligent distribution box according to claim 6, characterized in that, The specific method for obtaining the anomaly recovery index of the anomaly region is as follows: For any monitoring point in any abnormal region, obtain several target time periods corresponding to the abnormal time period of the monitoring point in the abnormal region, as several abnormal target time periods of the monitoring point in the abnormal region. Obtain the ratio of the sum of the time lengths of all the abnormal target time periods to the time length of the abnormal time period and use it as the third ratio. The ratio of the probability of the sudden change in the operating state of the monitoring point in the first abnormal target time period to the probability of the sudden change in the operating state in the last abnormal target time period in all the abnormal target time periods, and the result obtained by comparing it with the third ratio, is used as the abnormal fall-off index of the monitoring point. Obtain the center point of the abnormal region at the first moment and the center point at the last moment during the abnormal period, and denote them as the starting center point and the ending center point, respectively. Use the vector obtained by pointing the starting center point to the ending center point as the abnormal transmission vector of the abnormal region. Obtain the ratio of the area of ​​the local abnormal region corresponding to the next moment to the area of ​​the local abnormal region corresponding to the previous moment in any adjacent moment of the abnormal region. Based on the sum of the ratios obtained at all adjacent time points, the mean of the anomaly fall-off index of all monitoring points in the anomaly region, and the magnitude of the anomaly propagation vector, the anomaly recovery index of the anomaly region is obtained; the anomaly recovery index is positively correlated with the sum of the obtained ratios and the mean of the anomaly fall-off index, and negatively correlated with the magnitude.

8. The intelligent control method for a modular intelligent distribution box according to claim 1, characterized in that, The specific methods for obtaining the true anomaly probability of each anomaly region are as follows: For any local abnormal region at any time during an abnormal period of any abnormal region, the abnormal transmission vector of the abnormal region is assigned to the local abnormal region at that time. The cosine similarity between the abnormal transmission vector of the local abnormal region at that time and the abnormal transmission vectors of all other local abnormal regions is obtained. The reciprocal of the distance between the center points of two local abnormal regions is used as the weight to calculate the weighted average of all cosine similarities. The result is used as the reference factor for the position change of the abnormal region at that time during the abnormal period. Based on the anomaly recovery index of the abnormal region and the mean of the location change reference factor at all times during the abnormal period, the true anomaly probability of the abnormal region is obtained. The true anomaly probability is negatively correlated with both the anomaly recovery index and the mean of the location change reference factor. The true anomaly probability of the abnormal region is assigned to the corresponding local abnormal region at each time step in the abnormal region, thereby obtaining the true anomaly probability of each local abnormal region at each time step.

9. The intelligent control method for a modular intelligent distribution box according to claim 1, characterized in that, The specific method for obtaining the load limitation index of each local abnormal region at the current time includes: For any local anomaly region at the current moment, obtain the ratio of the true anomaly probability of the local anomaly region at the current moment to the average of the true anomaly probabilities of the local anomaly region at all historical moments. Based on the ratio, the number of monitoring points in the local anomaly area at the current time, and the average probability of sudden changes in the operating status of each monitoring point at the current time, the load limitation index of the local anomaly area at the current time is obtained.

10. A modular intelligent distribution box, comprising a distribution box, a monitoring unit, a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent control method for a modular intelligent distribution box as described in any one of claims 1-9, thereby performing intelligent control of the distribution box.

Images