Multi-branch current and voltage fusion monitoring method and system for low-voltage cable branch box

Abstract

The application discloses a multi-branch current and voltage fusion monitoring method and system for a low-voltage cable branch box, which comprises the following steps: synchronously sampling the branch currents and the busbar voltage of the same incoming busbar, calculating the voltage increment and the branch current increment; screening the effective time of the voltage increment meeting the threshold condition, and calculating the increment response ratio between any two branches; in the baseline learning period, the average value and the standard deviation of the increment response ratio of each branch are calculated to form an association baseline; in the monitoring stage, the standardized deviation is calculated according to the association baseline, and the association deviation scores of each branch are obtained to determine the suspected branch; then, the first-order difference of the increment response ratio time sequence in the backtracking window of the suspected branch is carried out, and the load switching and the poor contact are distinguished according to whether there is a step feature, and an alarm is output when the poor contact is confirmed. The application can improve the accuracy of early poor contact identification of the low-voltage cable branch box.

Description

Technical Field

[0001] This application relates to the field of cable branch box monitoring technology, specifically to a method and system for fusion monitoring of multi-branch current and voltage in low-voltage cable branch boxes. Background Technology

[0002] Low-voltage cable distribution boxes are widely used in low-voltage power supply systems such as communication base stations and communication equipment rooms. They are used to branch the main low-voltage cable to multiple branch circuits. Each branch is usually connected in parallel to the same incoming busbar and supplies power to the main communication equipment, air conditioning, lighting and other loads respectively. Since the distribution boxes are mostly installed outdoors or in unattended locations, the operating environment is greatly affected by temperature and humidity, load fluctuations and long-term energized operation. The contact resistance of the wiring terminals, knife switch contacts and other connection parts is prone to gradual increase due to oxidation, corrosion and loose crimping.

[0003] The voltage drop and current changes caused by poor contact in the early stages are usually small and difficult to reach the conventional overvoltage, undervoltage or overcurrent alarm thresholds. The existing monitoring method of setting thresholds independently for each branch is not easy to identify in the early stage of the fault. If the state continues to develop, it may cause local heating, terminal burning and even affect power supply safety under high current conditions. At the same time, loads such as air conditioning, lighting and communication equipment in low-voltage branch circuits are frequently switched on, off or switched on, and the normal current changes caused by them are easily confused with abnormal contact conditions, which further reduces the accuracy of alarms. Summary of the Invention

[0004] This application provides a method and system for fusion monitoring of multi-branch current and voltage in a low-voltage cable branch box, in order to at least solve some of the technical problems existing in the related technologies described above.

[0005] According to a first aspect of the embodiments of this application, a method for fusion monitoring of multi-branch current and voltage in a low-voltage cable branch box is provided, comprising: Synchronously sample the current and busbar voltage of each branch sharing the same incoming busbar in the branch box, and calculate the voltage increment of the busbar voltage relative to the previous sampling period and the current increment of each branch current relative to the previous sampling period. The time when the absolute value of the voltage increment is greater than the minimum effective threshold is selected from the sampling time. The incremental response ratio of any two branches at each effective time is calculated. The incremental response ratio is the quotient obtained by dividing the current increment of one branch by the current increment of the other branch. During the baseline learning period, the mean and standard deviation of the incremental response ratio of each branch pair at all effective times are calculated, and the mean and standard deviation are used as the associated baseline of the branch pair. During the monitoring phase, the standardized deviation of each branch pair is calculated. For each branch, the absolute values ​​of the standardized deviations of all branch pairs containing that branch are summed to obtain the association deviation score of that branch. Branches whose association deviation scores exceed a preset multiple of the median of the association deviation scores of all branches are marked as suspected branches. Perform first-order difference on the time series of incremental response ratios of the suspected branch within the backtracking window. Determine the cause of deviation based on whether there is a step feature in the difference series. If a step feature is present, cancel the suspected label and update the associated baseline. If no step feature is present, confirm that the branch has a poor contact and output an alarm.

[0006] As an optional approach, the minimum effective threshold is determined based on the busbar voltage sampling resolution, and its value is a preset integer multiple of the sampling resolution.

[0007] As an optional approach, when calculating the incremental response ratio, if the absolute value of the current increment of the branch used as the denominator is less than a preset integer multiple of the minimum effective resolution of the current transformer of that branch, the incremental response ratio with that branch as the denominator is not calculated.

[0008] As an optional approach, the calculation of the standardized deviation of each branch pair includes: subtracting the mean of the corresponding associated baseline from the current incremental response ratio of each branch pair and then dividing by the standard deviation of the corresponding associated baseline to obtain the standardized deviation of each branch pair, wherein the standardized deviation is taken as the moving average of the standardized deviations of the most recent preset number of valid times.

[0009] As an optional approach, determining whether the difference sequence has a step feature includes: taking the point with the largest absolute value in the difference sequence; if the absolute value of the point exceeds a first preset multiple of the standard deviation of the difference sequence, and the absolute values ​​of the difference mean within a preset window before and after the point are both lower than a second preset multiple of the standard deviation of the difference sequence, then it is determined that a step feature exists, wherein the first preset multiple is greater than the second preset multiple.

[0010] As an optional approach, after canceling the suspected branch and updating the associated baseline, if multiple step features exist within the backtracking window, the associated deviation score of the branch is recalculated after each update of the associated baseline. If the branch still meets the conditions for being marked as a suspected branch, then the poor contact of the branch is confirmed and an alarm is output.

[0011] As an optional approach, after confirming that a poor contact has occurred in the branch, the median of the ratio of the busbar voltage increment to the branch current increment at each effective moment within the most recent preset time period is taken as the current equivalent impedance estimate. The equivalent impedance baseline value obtained in the same way during the baseline learning period is subtracted from the current equivalent impedance estimate, and the difference is taken as the contact resistance increment estimate of the branch. The contact resistance increment estimate is output along with the alarm.

[0012] As an optional approach, during the monitoring phase, dynamically updating the associated baseline at a preset period includes: within a sliding statistical window of the most recent preset number of days, excluding data identified as transition segments before and after load switching events and data of branch pairs marked as suspected branch pairs, and recalculating the mean and standard deviation using the incremental response ratio of the remaining valid time periods to replace the original associated baseline.

[0013] As an optional approach, when the number of branches marked as suspected exceeds a preset proportion of the total number of branches, the coefficient of variation of the correlation deviation score of all branches is calculated. If the coefficient of variation is lower than a preset value, an abnormality prompt is output on the incoming line side. If the coefficient of variation is not lower than the preset value, the individual suspected branches are located according to the correlation deviation score.

[0014] According to a second aspect of the embodiments of this application, a multi-branch current and voltage fusion monitoring system for a low-voltage cable branch box is also provided, comprising: The synchronous sampling and calculation module is used to synchronously sample the current and busbar voltage of each branch sharing the same incoming busbar in the branch box, and calculate the voltage increment of the busbar voltage relative to the previous sampling period and the current increment of each branch current relative to the previous sampling period. The effective time screening and ratio calculation module is used to screen the time when the absolute value of the voltage increment is greater than the minimum effective threshold from each sampling time as the effective time, and calculate the incremental response ratio of the branch formed by any two branches at each effective time. The incremental response ratio is the quotient obtained by dividing the current increment of one branch by the current increment of the other branch. The associated baseline establishment module is used to calculate the mean and standard deviation of the incremental response ratio of each branch pair at all effective times during the baseline learning period, and use the mean and standard deviation as the associated baseline of the branch pair. The suspected branch identification module is used to calculate the standardized deviation of each branch pair during the monitoring phase. For each branch, the absolute values ​​of the standardized deviations of all branch pairs containing that branch are summed to obtain the association deviation score of that branch. Branches whose association deviation scores exceed a preset multiple of the median of the association deviation scores of all branches are marked as suspected branches. The deviation cause determination and alarm module is used to perform first-order difference on the incremental response ratio time series of the suspected branch within the backtracking window. The deviation cause is determined based on whether there is a step feature in the difference sequence. If a step feature is present, the suspected label is canceled and the associated baseline is updated. If no step feature is present, the poor contact of the branch is confirmed and an alarm is output.

[0015] This application utilizes the parallel structure of multiple branches sharing the same incoming busbar within a low-voltage cable branch box to synchronously analyze the busbar voltage increment and the current increment of each branch, establishing a correlation baseline for the incremental response ratio between branches. Abnormal branches are located through correlation deviation scoring. Compared to monitoring methods relying solely on single-branch voltage and current thresholds, this approach can identify early contact defects before contact resistance causes significant overvoltage, undervoltage, or overcurrent. By performing first-order differential judgment on the ratio sequence within the suspected branch's backtracking window, it can distinguish between abrupt changes caused by normal load input / output and gradual deviations caused by a slow increase in contact resistance, thereby reducing false alarms. Furthermore, dynamically updating the correlation baseline and outputting estimated contact resistance increments allows the monitoring results to adapt to long-term load changes and provides a basis for on-site maintenance to determine fault location and severity, thus improving the safety and maintenance efficiency of low-voltage branch power supply systems.

[0016] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Furthermore, no embodiment in this disclosure is required to achieve all the effects described above. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0018] Figure 1 A flowchart illustrating a method for fusion monitoring of multi-branch current and voltage in a low-voltage cable branch box, as provided in this embodiment of the disclosure.

[0019] Figure 2 This is a schematic diagram illustrating the calculation process of the incremental response ratio of a branch pair provided in an embodiment of this disclosure.

[0020] Figure 3 This is a schematic diagram of the abnormal branch location process provided in the embodiments of this disclosure.

[0021] Figure 4 This is a schematic diagram of the suspected branch verification process provided in an embodiment of this disclosure.

[0022] Figure 5 This is a schematic block diagram of a multi-branch current and voltage fusion monitoring system for a low-voltage cable branch box provided in an embodiment of this disclosure. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0024] In the power supply system of communication base stations and communication equipment rooms, low-voltage cable branch boxes serve the function of distributing the main low-voltage cables to multiple branch circuits. Each branch circuit supplies power to different types of loads such as main communication equipment, air conditioning, and lighting. The internal topology of the branch box includes a pair of incoming busbars as a common combiner node, with all branches connected in parallel from this busbar through their respective terminals and disconnectors. The number of branches is determined based on the actual project requirements; for example, it can be configured with 8, 12, or 16 branches. Branch boxes are mostly deployed outdoors or in unattended environments where temperature and humidity fluctuate significantly.

[0025] During long-term operation, the contact resistance at the connection points of each branch may slowly increase due to oxidation, corrosion and other factors. The voltage drop and current deviation caused by this early poor contact are extremely small and far below the conventional overvoltage, undervoltage or overcurrent alarm thresholds. The monitoring method of setting thresholds for each branch independently is difficult to identify it in the early stage. However, once the contact resistance continues to increase, the local heating under high current conditions can lead to terminal burn-out or even fire.

[0026] Therefore, this method utilizes the parallel topology of shared incoming busbars within the same branch box to establish a quantitative correlation between branches based on their dynamic response to busbar voltage fluctuations. Then, by detecting deviations from this correlation, early contact failures in specific branches can be identified. The following assumes that the branch box has a total of... Branch roads, each branch road numbered This method is implemented at the hardware level based on the existing conventional current and voltage acquisition devices in the branch box. A current transformer is configured at the outlet of each branch to acquire the branch current, and a voltage acquisition module is configured at the incoming busbar to acquire the busbar voltage. The sampling clocks of each acquisition channel are kept synchronized, and the entire processing can be completed in the processing unit of the monitoring terminal.

[0027] The implementation process of the method described in this application will be described in detail below with reference to specific embodiments. It should be noted that this embodiment is only used to explain this application and is not intended to limit the scope of protection of this application. Conventional adjustments or substitutions of each step by those skilled in the art without departing from the concept of this application should be included in the scope of protection of this application.

[0028] Please see Figure 1 , Figure 1 This is a flowchart of a multi-branch current and voltage fusion monitoring method for a low-voltage cable branch box according to an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes steps S1-S5: In step S1, the current and busbar voltage of each branch sharing the same incoming busbar in the branch box are synchronously sampled, and the voltage increment of the busbar voltage and the current increment of each branch current relative to the previous sampling period are calculated respectively.

[0029] In one embodiment, the processing unit of the monitoring terminal performs sampling on the busbar voltage and all [other parameters] at a fixed sampling period. The current of each branch is sampled synchronously; the sampling period can be configured according to the actual situation, for example, it can be set to 1 second; in each sampling period, the processing unit reads the bus voltage value and the current value of each branch at the current moment, and then subtracts the bus voltage value of the previous sampling period from the bus voltage value at the current moment. The difference is the voltage increment at that moment; similarly, for each branch, the current value at the current moment is subtracted from the current value of the previous sampling period. The difference is the current increment of that branch at that moment.

[0030] The reason for using incremental values ​​is that the load capacity and type of each branch are different, and the reference level of the absolute current value varies greatly. Directly comparing the absolute values ​​makes it difficult to reflect the difference in the response of each branch to the bus voltage fluctuation. Incremental values ​​only retain the dynamic response information of each branch as the voltage changes, and remove the difference in the load reference level from the data, providing a unified and comparable data basis for establishing the correlation between branches in the future.

[0031] In step S2, the time when the absolute value of the voltage increment is greater than the minimum effective threshold is selected from each sampling time as the effective time. The incremental response ratio of any two branches at each effective time is calculated, wherein the incremental response ratio is the quotient obtained by dividing the current increment of one branch by the current increment of the other branch.

[0032] Please see Figure 2 , Figure 2 A schematic diagram of the branch-to-incremental response ratio calculation process provided in an embodiment of this disclosure is shown, as follows: Figure 2 As shown, in S201, a minimum effective threshold is determined, and the time when the absolute value of the voltage increment is greater than the minimum effective threshold is selected from each sampling time as the effective time.

[0033] Understandably, not every incremental data point at every sampling moment is suitable for subsequent correlation analysis. When the absolute value of the busbar voltage increment is too small, the current increment of each branch mainly consists of measurement noise, and the ratio calculated based on this will fluctuate wildly and have no physical meaning. Therefore, the processing unit first determines whether the absolute value of the voltage increment is greater than a minimum effective threshold at each sampling moment. Only when this condition is met is the moment recognized as a valid moment and included in subsequent processing.

[0034] The value of this minimum effective threshold is directly related to the sampling resolution of the bus voltage acquisition module. Specifically, it can be set to a preset integer multiple of the sampling resolution, such as 5 to 10 times the sampling resolution. The basis for this setting is that the voltage increment should be significantly higher than the inherent quantization noise level of the acquisition system in order to ensure that the result of the subsequent ratio calculation reflects the real physical response rather than the quantization error. This threshold is determined by the hardware accuracy index and the preset integer multiple, and is a data validity threshold set based on the hardware resolution.

[0035] In S202, at each valid time, the processing unit processes any two branches. and Calculate the incremental response ratio. During the baseline learning period, monitoring phase, and associated baseline update phase, a consistent numerator and denominator order should be used for the same branch pair, for example, fixed according to the branch number order. Sometimes by side road The current increment is in the molecule and branch. The current increment is the denominator; other pre-fixed sequences can also be used, but the sequence of the same branch pair must not be changed in each stage; the branch The current increment at that moment divided by the branch The current increment at that moment, the quotient obtained is the branch. and branch road The incremental response ratio at that moment.

[0036] Since all branches are connected in parallel to the same busbar, when the busbar voltage fluctuates, the change in current of each branch depends on the equivalent admittance of that branch, which is the reciprocal of the equivalent impedance. This equivalent impedance is composed of the load impedance and the contact resistance at the connection point in series. When the contact resistance is normal and extremely small, the load impedance is dominant. The physical essence of the ratio of the incremental response of the two branches is the ratio of the equivalent admittance of the two branches. When the contact state and load type of the two branches do not change, this ratio remains stable for a period of time.

[0037] Among them, when the branch is the denominator If the absolute value of the current increment is too small, that is, lower than a preset integer multiple of the minimum effective resolution of the current transformer in that branch, then the branch... The ratio with a small denominator is not calculated because a small denominator makes the ratio extremely sensitive to minute noise, and the calculated ratio is not of any reference value.

[0038] for A branch road can form For each branch pair, a set of incremental response ratios corresponding to these branch pairs is generated at each effective time. At each effective time, the incremental response ratio is calculated and recorded only for branch pairs whose absolute values ​​of branch current increments in the denominator satisfy the aforementioned minimum effective resolution condition. Branch pairs that do not satisfy this condition are not included in the ratio sequence of that branch pair at that effective time.

[0039] In step S3, during the baseline learning period, the mean and standard deviation of the incremental response ratio of each branch pair at all effective times are calculated, and the mean and standard deviation are used as the associated baseline of the branch pair.

[0040] In one embodiment, after the system is powered on, it enters a baseline learning period. During this period, the processing unit continuously acquires incremental data and calculates the incremental response ratio of each branch pair at each effective time according to the aforementioned steps. At the same time, these ratios are stored separately for each branch pair. At the end of the baseline learning period, the processing unit calculates the mean and standard deviation of the incremental response ratio sequence that has been calculated and recorded during the learning period for each branch pair, and uses the mean and standard deviation together as the associated baseline of the branch pair.

[0041] For any branch pair, if the number of incremental response ratios calculated and recorded during the baseline learning period is lower than the preset minimum baseline sample size, then the associated baseline for that branch pair will not be established temporarily, and the baseline learning period will be extended or await subsequent valid data to supplement it. The preset minimum baseline sample size can be set according to the minimum sample size required for the statistical mean and standard deviation, for example, no less than 30 valid ratios. If the statistically obtained standard deviation is 0 or lower than the minimum standard deviation threshold, then the standard deviation of that branch pair will be taken as the minimum standard deviation threshold. The minimum standard deviation threshold is determined by converting the incremental response ratio based on the sampling resolution of the busbar voltage acquisition module and the branch current transformer.

[0042] The mean of the associated baseline represents the typical level of the equivalent admittance ratio of the two branches under normal contact conditions; the standard deviation reflects the fluctuation range of this ratio under normal operating conditions, which is affected by factors such as load fluctuation and measurement noise.

[0043] In one embodiment, the baseline learning period should cover at least one complete load day cycle for the branch box, ensuring that the statistical results encompass typical operating conditions such as daytime peaks and nighttime troughs; for example, it can be configured to be 48 to 96 hours. Furthermore, the learning period should be scheduled after initial system deployment or maintenance and repair, and after normal contact status confirmation, to ensure that the baseline reflects the inter-branch relationships under normal contact conditions. In some embodiments, after the branch box has undergone maintenance and repair and the contact status has returned to normal, the baseline learning process can be retried to update the associated baseline.

[0044] In step S4, during the monitoring phase, the standardized deviation of each branch pair is calculated. For each branch, the absolute values ​​of the standardized deviations of all branch pairs containing that branch are summed to obtain the association deviation score of that branch. Branches whose association deviation scores exceed a preset multiple of the median of the association deviation scores of all branches are marked as suspected branches.

[0045] In one embodiment, after the baseline learning period, the system enters the normal monitoring phase. The processing unit continuously acquires incremental data, filters valid moments, and calculates the incremental response ratio of each branch pair. Based on this, it performs standardized deviation calculation and abnormal branch location. The impedance change of a single branch caused by early contact failure is extremely small, and the absolute value deviation is easily drowned out by measurement noise and normal load fluctuations. If the sensitivity is simply increased by lowering the threshold, independent judgment of each branch can easily introduce false alarms, making it difficult to balance sensitivity and false alarm rate. Therefore, this embodiment transforms the detection object from the absolute value of a single branch to the relative deviation between branches, and uses the structural distribution characteristics of the fault in multiple branch pairs for location, achieving high-sensitivity identification of early contact failure.

[0046] Please see Figure 3 , Figure 3 A schematic diagram of the abnormal branch location process provided in the embodiments of this disclosure is shown, such as... Figure 3 As shown in S301, the standardized deviation of each branch pair is calculated during the monitoring phase.

[0047] For each branch pair, the processing unit compares the current incremental response ratio of the branch pair with its associated baseline at each effective time: subtracts the mean of the associated baseline of the branch pair from the current incremental response ratio, and divides the difference by the standard deviation of the associated baseline of the branch pair, where the standard deviation includes the standard deviation after processing the minimum standard deviation threshold determined in the aforementioned steps, and finally obtains the standardized deviation of the branch pair at the current time.

[0048] The standardized deviation measures the degree of deviation of the current ratio from the normal level, using the baseline standard deviation as a scale. The larger the absolute value, the more the equivalent admittance of at least one branch in the branch pair has deviated from the normal state. Since the standard deviation itself already contains information about the ratio fluctuation under normal operating conditions, the deviations between different branch pairs are comparable after standardization and are no longer affected by the differences in admittance ratios of each branch pair to the baseline level.

[0049] The change in current in a single branch caused by an increase of a few milliohms in contact resistance may only be tens of milliamperes in absolute terms, far smaller than the accuracy range of the current transformer and the normal fluctuation range of the load itself, and almost imperceptible when judging each branch independently. However, this tiny impedance change will cause the admittance ratio between the faulty branch and every other normal branch to shift, while the admittance ratio between normal branches remains unaffected. The standardized deviation quantifies this relative shift using the baseline standard deviation as a scale, amplifying the weak signal that was originally submerged in noise to a identifiable level at the relational level.

[0050] In practical implementation, the standardized deviation at a single valid moment may fluctuate significantly due to transient noise. Therefore, a moving average of the standardized deviations over several recent valid moments can be taken, and this moving average can be used as the standardized deviation of the branch relative to the current usage. The length of the moving average window is configurable, for example, it can be set to the most recent 10 to 30 valid moments. A longer window provides better smoothing for transient noise, but the response speed to rapidly changing contact defects is correspondingly reduced. In some embodiments, a suitable window length can be selected within the above range based on the noise level of the power supply environment where the branch box is located and the required inspection cycle.

[0051] In S302, the absolute values ​​of the standardized deviations of all branch pairs containing the branch are summed to obtain the association deviation score of the branch. Branches whose association deviation scores exceed a preset multiple of the median of the association deviation scores of all branches are marked as suspected branches.

[0052] The standardized deviation of each branch pair reflects whether the relationship between the two branches deviates from normal. However, based solely on the abnormal deviation of a single branch pair, it is not possible to directly determine which branch in that pair is experiencing a problem. For example, if the incremental response ratio of branch 3 and branch 7 deviates, it could be due to an increase in the contact resistance of branch 3 or branch 7. The two situations are indistinguishable in terms of the deviation of the branch pair. Therefore, this embodiment utilizes the structural consistency characteristics of the fault in the relationship between multiple branch pairs to solve the problem of difficult localization.

[0053] Specifically, if branch road Poor contact leads to an increase in its equivalent impedance, then the branch With branch roads All branch pairs formed by the branches other than those branches, i.e., branches and branch road The branch pairs formed, among which and (common (each), its incremental response ratio will be affected by the branch. The shift occurs due to changes in admittance, and the corresponding standardized deviations will all show anomalies. Additionally, this excludes branches. Any pair of branches, such as branch 2 and branch 5, are not branches. At that time, the incremental response ratio of the branch pair formed by branch 2 and branch 5 is not affected by the branch. Due to the influence of [the factors], the standardized deviation remained within the normal range.

[0054] This structural distribution characteristic, where anomalies are concentrated in the same branch, is an inherent manifestation of single-branch faults in parallel topologies. It does not depend on the severity of the fault; as long as the impedance change of the faulty branch is sufficient to be identified at the level of standardized deviation, this structural characteristic will appear.

[0055] Based on the above principles, the processing unit processes each branch. The correlation deviation score is calculated by traversing all branches containing the path. For each pair of branches, sum the absolute values ​​of their standardized deviations one by one; the sum is the branch. Correlation deviation score; including branches The branch road is shared One, that is, a branch road Each branch is paired with other branches besides itself, therefore the association deviation score for each branch is... The sum of the absolute values ​​of standardized deviations. This score reflects the overall branch... The degree of deviation in all branches associated with it: if the branch If the contact is normal, the standardized deviations of each associated branch are small, and the cumulative score is also low; if the branch If there is poor contact, all associated branches will have higher deviations from the standardization, and the cumulative score will be much higher than that of normal branches.

[0056] The processing unit calculates all After each branch has its own correlation deviation score, the median of all scores is taken as the reference benchmark. If the correlation deviation score of a branch exceeds a preset multiple of the median, the processing unit marks the branch as a suspected branch. The preset multiple can be configured according to the actual situation, for example, it can be set to 3 times. The larger the multiple, the stricter the judgment and the fewer false alarms, but the sensitivity to minor faults will be reduced accordingly.

[0057] In step S5, the time series of incremental response ratios of the suspected branch within the backtracking window is subjected to first-order difference. The cause of deviation is determined based on whether there is a step feature in the difference series. If a step feature exists, the suspected branch is removed and the associated baseline is updated. If no step feature exists, the branch is confirmed to have poor contact and an alarm is output.

[0058] In actual communication scenarios, the loads on each branch in a low-voltage cable branch box are switched on or off as needed for normal operation. Examples include air conditioning compressors starting and stopping according to temperature control signals, lighting switching on and off according to schedules, and communication equipment switching between primary and backup modes. Air conditioning equipment in communication base stations is particularly active, potentially starting and stopping multiple times per hour during hot summer months. Each load switch causes a sudden change in the equivalent impedance of that branch; for example, the equivalent impedance decreases when the load is switched on and increases when the load is switched off. This leads to a step change in the incremental response ratio between that branch and other branches. This step change also increases the normalization deviation, triggering a suspicious flag, but the cause is normal load switching rather than poor contact. Without distinguishing between the two, frequent load switching in the low-voltage cable branch box will generate numerous false alarms, drowning out genuine poor contact alarms. Therefore, this embodiment uses the different time-series characteristics of the two types of events in the incremental response ratio time series to differentiate them.

[0059] Specifically, the increase in contact resistance is a physical process in which the oxide layer on the metal surface gradually grows and the contact area gradually decreases. Its time scale is usually several days to several weeks. It is reflected in the incremental response ratio time series as a continuous, consistent, and slow shift, or as the magnitude of the deviation from the baseline gradually increases over a longer period of time.

[0060] Load switching is an instantaneous event, with a stable ratio range before and after the switch. A jump occurs at the moment of switching. If we perform a first-order difference on the incremental response ratio time series—that is, calculate the change in ratio between adjacent effective moments—load switching appears as a single or a few pulse points with significantly larger absolute values. The difference before and after these pulses quickly returns to near zero, indicating that the ratio has stabilized at a new level after the jump and no longer changes continuously. In contrast, poor contact in the difference series manifests as multiple small, continuous shifts in the same direction, or a deviation that does not tend to return to near zero over a longer period.

[0061] Load switching changes the load impedance itself. After a switch is completed, the new load impedance stabilizes immediately, and the equivalent admittance ratio jumps to the new value and then stops changing. Poor contact, on the other hand, changes the contact resistance at the connection point. Its degradation process is continuous and gradual, and the deviation of the equivalent admittance ratio will not stop and stabilize at a certain moment.

[0062] Please see Figure 4 , Figure 4 A schematic diagram of the suspected branch verification process provided in this embodiment of the present disclosure is shown, such as... Figure 4 As shown, in S401, the time series of the incremental response ratio of the suspected branch within the backtracking window is subjected to first-order difference.

[0063] Based on the above principles, when a certain branch road After a branch is marked as a suspect, the processing unit extracts the time series of incremental response ratios associated with that branch within the most recent backtracking window. The duration of the backtracking window can be configured according to actual conditions, for example, set to the most recent 24 hours. The extracted time series refers to the time series containing the branch within all valid moments within the backtracking window. The incremental response ratio sequence of each branch pair; the processing unit for each branch The branch pairs are subjected to first-order differencing of their incremental response ratio sequences, and the branches are further differentiated based on whether their differencing sequences exhibit step characteristics. The cause of the deviation is determined; to simplify the process, the average of the standardized absolute values ​​of the deviations of these branches at each valid time can be taken to form a line reflecting the deviations of the branches. The overall deviation level changes over time.

[0064] The processing unit calculates the first-order difference of the above incremental response ratio time series, that is, by subtracting the incremental response ratio of the previous valid time from the incremental response ratio of each valid time, to form a difference sequence. Each value in the difference sequence reflects the branch. The magnitude and direction of the change in the incremental response ratio of the involved branch pairs between two adjacent effective times.

[0065] In S402, the processing unit determines whether the difference sequence exhibits a step characteristic according to preset rules. Specifically, the preset rules include: first, finding the point with the largest absolute value in the difference sequence, and then checking two conditions: first, whether the absolute value of this point exceeds a first preset multiple of the standard deviation of the difference sequence; second, whether the absolute values ​​of the difference mean within a preset window before this point and the absolute values ​​of the difference mean within a preset window after this point are both lower than a second preset multiple of the standard deviation of the difference sequence; the first preset multiple is configurable, for example, set to 5 times; the second preset multiple is configurable, for example, set to 1.5 times; the lengths of the preset windows before and after are configurable, for example, taking 10 valid time points each.

[0066] The first condition determines whether there is a sudden change in the difference sequence with an amplitude far exceeding normal fluctuations. Only when the magnitude of the change is sufficiently significant is it meaningful to discuss its cause. The second condition confirms that the differences before and after the change point are at low levels, meaning the ratio is stable before the change and quickly stabilizes afterward. When both conditions are met, the difference sequence conforms to the step characteristic of settling after a jump. If the first condition is met but the second condition is not, for example, if the difference continues to deviate from zero after the change point, it indicates that the ratio has not stabilized after the jump and does not conform to the step characteristic. If the first condition is not met, it indicates that there is no significant change point in the difference sequence, and the deviation is gradual rather than sudden, which also does not conform to the step characteristic.

[0067] In S403, based on the determination result, the processing unit performs subsequent operations in two ways. Specifically, if a step characteristic is determined, it indicates that the ratio deviation of the suspected branch is caused by load switching, and the processing unit cancels the suspected label of the branch. At the same time, the mean and standard deviation of the incremental response ratios of each branch pair involved in the branch are recalculated using the steady-state data after the load switching (i.e., the effective time data after the step point) to update the corresponding correlation baseline. The purpose of the update is to adapt subsequent monitoring to the new operating state after the load switching and avoid repeated triggering of the suspected label during the same load switching.

[0068] If no step characteristic is found, it means that the deviation of the incremental response ratio of the branch is a gradual process caused by the slow increase of contact resistance. The processing unit confirms that the branch has poor contact and outputs an alarm message. The alarm message includes the number of the suspected branch and the associated deviation score of the branch.

[0069] In some embodiments, a branch may have experienced multiple load switchings within the backtracking window period, such as an air conditioner starting and stopping multiple times in a short period of time. At this time, multiple abrupt change points satisfying the step feature conditions may appear in the differential sequence of one or more branch pairs containing the branch. The processing unit performs baseline updates for each identified step feature in sequence: updates the associated baseline with the steady-state segment data after each step point, and then recalculates the standardized deviation and associated deviation score of the branch based on the updated baseline. If the associated deviation score of the branch still exceeds a preset multiple of the median score of all branches after all baseline updates, it indicates that after excluding the influence of all identifiable load switching, the association relationship between the branch and other branches still has abnormal deviations. The processing unit confirms it as a poor contact and outputs an alarm.

[0070] This ensures that in communication power supply scenarios with frequent load switching, each normal load switch can be identified and eliminated one by one, while the gradual deviation caused by poor contact will not be masked by repeated baseline updates.

[0071] In some embodiments, optionally, after confirming that a branch has poor contact, the processing unit further estimates the increase in the contact resistance of the branch, providing a quantitative reference for maintenance personnel to judge the severity of the fault.

[0072] Specifically, the processing unit extracts the busbar voltage increment and the branch current increment at each effective moment within the most recent preset time period. The most recent preset time period can be configured to be the most recent 24 hours to 72 hours, and should include no less than a preset minimum number of available effective moments. If the number of available effective moments is insufficient, the preset time period is extended until the preset minimum number is met. For effective moments where the absolute value of the branch current increment is less than a preset integer multiple of the minimum effective resolution of the branch current transformer, the equivalent impedance estimate for that moment is not calculated. Instead, the busbar voltage increment is divided by the branch current increment at each available effective moment to obtain the equivalent impedance estimate for that moment. Then, the median of all available equivalent impedance estimates within the preset time period is taken as the current equivalent impedance estimate.

[0073] The current equivalent impedance estimate is subtracted from the equivalent impedance baseline value obtained in the same way during the baseline learning period. That is, the median of the ratio of the voltage increment to the current increment of the branch at each effective moment that satisfies the validity condition of the branch current increment during the baseline learning period is obtained. The difference is the incremental estimate of the contact resistance of the branch. The processing unit outputs the incremental estimate of the contact resistance along with the alarm information.

[0074] If the branch has undergone load switching and updated the associated baseline before the alarm, the equivalent impedance baseline value should be recalculated using the steady-state data after the most recent baseline update to eliminate the interference of load changes on the estimation results.

[0075] In one embodiment, during long-term operation, the load configuration of each branch may change due to the addition, removal or adjustment of equipment. The associated baseline established during the initial baseline learning period may not always be applicable. In order to maintain the matching between the associated baseline and the actual operating conditions, the processing unit dynamically updates the associated baseline at a preset cycle during the monitoring phase.

[0076] Specifically, the processing unit maintains a sliding statistical window, which can be configured to cover the most recent 7 to 30 days. When collecting incremental response ratio data for all valid moments within this window, two types of data are excluded: first, data already identified as the transition segment before and after a load switching event; and second, data of branch pairs still marked as suspected branch pairs. The transition segment before and after a load switching event refers to the valid moments corresponding to a preset window length used for step feature determination, taken forward and backward from the identified step point. When the on-site load switching duration is long, the transition segment range can be expanded by a preset integer multiple of the preset window length. The purpose of excluding these two types of data is to prevent abnormal or transitional state data from being mixed into the baseline statistics and causing baseline distortion. The mean and standard deviation of the incremental response ratio of each branch pair are recalculated using the remaining valid moment data after exclusion, and the resulting data replaces the original associated baseline for deviation detection.

[0077] The update cycle can be configured according to actual conditions, such as once every 24 hours. When the number of valid moments remaining after exclusion within the sliding statistical window is lower than the preset minimum value, it indicates that the available data within the window is insufficient to guarantee statistical reliability. The processing unit will continue to use the most recently updated associated baseline and will not perform this update, while outputting a message indicating insufficient data. The preset minimum value can be set according to the minimum sample size required for the mean and standard deviation statistics, such as no less than 30 valid ratios. The associated baseline update will only be performed on the branch pair when the number of remaining valid ratios for each branch pair meets the preset minimum value.

[0078] In some embodiments, optionally, multiple branches may simultaneously meet the suspected marking criteria. There are two possible causes: either multiple branches genuinely have poor contact, or the incoming busbar itself has poor contact at its connection points, causing multiple branches to simultaneously show increased correlation deviation scores during the same monitoring period. The latter situation differs from the former. Specifically, incoming-side anomalies affect the operating status of the common incoming side of the branch box, potentially causing multiple branches to be marked as suspected branches simultaneously. In this case, it is not appropriate to directly interpret the increased score as poor contact in a single branch. If the correlation deviation scores of each branch increase relatively evenly, the processing unit uses this as one of the criteria for judging incoming-side anomaly alerts, rather than directly locating individual branches based on a single highest score.

[0079] To distinguish between these two situations, when the number of suspected branches exceeds a preset proportion of the total number of branches, the processing unit calculates the coefficient of variation of the correlation deviation scores of all branches, which is the quotient obtained by dividing the standard deviation of the correlation deviation scores of all branches by their mean. If the coefficient of variation is lower than the preset value, for example, it can be configured to 0.3, indicating that the scores of each branch are relatively evenly distributed and the differences between them are not significant. At this time, the processing unit outputs an abnormality prompt on the incoming line side, reminding maintenance personnel to focus on checking the connection status of the common incoming line side of the branch box and the incoming busbar. If the coefficient of variation is not lower than the preset value, it indicates that there are significant differences in the score distribution, and some branches still have scores much higher than other branches. The processing unit maintains the processing method of locating individual suspected branches according to the correlation deviation scores in the aforementioned steps.

[0080] According to the embodiments of this disclosure, by utilizing the multi-branch parallel topology association information contained in the existing current and voltage acquisition data, the detection sensitivity of early contact failure is improved from being limited by the measurement noise level of a single branch to the detection level of deviation in the relationship between branches. The specific faulty branch is determined by the structural consistency of the fault in the relationship between multiple branches, thereby reducing the false alarm rate. The first-order differential verification effectively eliminates the main interference source in the communication power supply scenario with frequent load switching, making the contact failure alarm highly reliable.

[0081] Please see Figure 5 , Figure 5 This is a structural block diagram of a multi-branch current and voltage fusion monitoring system for a low-voltage cable branch box provided in an embodiment of this application. Figure 5 As shown, the system includes: The synchronous sampling calculation module 501 is used to synchronously sample the current and busbar voltage of each branch sharing the same incoming busbar in the branch box, and calculate the voltage increment of the busbar voltage relative to the previous sampling period and the current increment of each branch current relative to the previous sampling period. The effective time screening and ratio calculation module 502 is used to screen the time when the absolute value of the voltage increment is greater than the minimum effective threshold from each sampling time as the effective time, and calculate the incremental response ratio of the branch formed by any two branches at each effective time, wherein the incremental response ratio is the quotient obtained by dividing the current increment of one branch by the current increment of the other branch. The associated baseline establishment module 503 is used to calculate the mean and standard deviation of the incremental response ratio of each branch pair at all effective times during the baseline learning period, and use the mean and standard deviation as the associated baseline of the branch pair. The suspected branch identification module 504 is used to calculate the standardized deviation of each branch pair during the monitoring phase. For each branch, the absolute values ​​of the standardized deviations of all branch pairs containing the branch are summed to obtain the association deviation score of the branch. Branches whose association deviation scores exceed a preset multiple of the median of the association deviation scores of all branches are marked as suspected branches. The deviation cause determination and alarm module 505 is used to perform first-order difference on the incremental response ratio time series of the suspected branch within the backtracking window, and determine the deviation cause based on whether there is a step feature in the difference series. If there is a step feature, the suspected mark is canceled and the associated baseline is updated. If there is no step feature, it is confirmed that the branch has poor contact and an alarm is output.

[0082] Each processing unit and / or module in the embodiments of this application can be implemented by an analog circuit that implements the functions described in the embodiments of this application, or by software that executes the functions described in the embodiments of this application.

[0083] In particular, the methods and / or embodiments in this application can be implemented as computer software programs. For example, the embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. When the computer program is executed by a processor, it performs the functions defined in the methods of this application.

[0084] Another embodiment of this application provides a storage medium storing computer program instructions thereon, which can be executed by a processor to implement the methods and / or technical solutions of any one or more embodiments of this application.

[0085] In the above embodiments, the descriptions of each embodiment have different focuses. Parts not described in detail in a certain embodiment can be referred to in the relevant descriptions of other embodiments. The above descriptions are merely preferred embodiments of this application and explanations of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to the technical solutions formed by specific combinations of the above technical features, but should also cover other technical solutions formed by arbitrary combinations of the above technical features or their equivalent features without departing from the inventive concept.

Claims

1. A method for fusion monitoring of multi-branch current and voltage in a low-voltage cable branch box, characterized in that, include: Synchronously sample the current and busbar voltage of each branch sharing the same incoming busbar in the branch box, and calculate the voltage increment of the busbar voltage relative to the previous sampling period and the current increment of each branch current relative to the previous sampling period. The time when the absolute value of the voltage increment is greater than the minimum effective threshold is selected from the sampling time. The incremental response ratio of any two branches at each effective time is calculated. The incremental response ratio is the quotient obtained by dividing the current increment of one branch by the current increment of the other branch. During the baseline learning period, the mean and standard deviation of the incremental response ratio of each branch pair at all effective times are calculated, and the mean and standard deviation are used as the associated baseline of the branch pair. During the monitoring phase, the standardized deviation of each branch pair is calculated. For each branch, the absolute values ​​of the standardized deviations of all branch pairs containing that branch are summed to obtain the association deviation score of that branch. Branches whose association deviation scores exceed a preset multiple of the median of the association deviation scores of all branches are marked as suspected branches. The incremental response ratio time series of the suspected branch within the backtracking window is subjected to first-order differencing. The cause of deviation is determined based on whether there is a step feature in the differencing sequence. If a step feature is present, the suspected branch is removed and the associated baseline is updated. If no step feature is present, the branch is confirmed to have poor contact and an alarm is output. Determining whether there is a step feature in the differencing sequence includes: taking the point with the largest absolute value in the differencing sequence. If the absolute value of this point exceeds a first preset multiple of the standard deviation of the differencing sequence, and the absolute values ​​of the mean of the differences within a preset window before and after this point are both lower than a second preset multiple of the standard deviation of the differencing sequence, then a step feature is determined to exist, wherein the first preset multiple is greater than the second preset multiple.

2. The method according to claim 1, characterized in that, The minimum effective threshold is determined based on the busbar voltage sampling resolution and is a preset integer multiple of that sampling resolution.

3. The method according to claim 1, characterized in that, When calculating the incremental response ratio, if the absolute value of the current increment of the branch used as the denominator is less than a preset integer multiple of the minimum effective resolution of the current transformer of that branch, the incremental response ratio with that branch as the denominator is not calculated.

4. The method according to claim 1, characterized in that, The calculation of the standardized deviation of each branch pair includes: subtracting the mean of the corresponding associated baseline from the current incremental response ratio of each branch pair and then dividing by the standard deviation of the corresponding associated baseline to obtain the standardized deviation of each branch pair, wherein the standardized deviation is the moving average of the standardized deviations of the most recent preset number of valid times.

5. The method according to claim 1, characterized in that, After canceling the suspected branch and updating the associated baseline, if there are multiple step features in the backtracking window, the associated deviation score of the branch is recalculated after each update of the associated baseline. If the branch still meets the conditions for being marked as a suspected branch, it is confirmed that the branch has poor contact and an alarm is output.

6. The method according to claim 1, characterized in that, After confirming that a poor contact has occurred in the branch, the median of the ratio of the busbar voltage increment to the branch current increment at each effective moment within the most recent preset time period is taken as the current equivalent impedance estimate. The equivalent impedance baseline value obtained in the same way during the baseline learning period is subtracted from the current equivalent impedance estimate, and the difference is taken as the contact resistance increment estimate of the branch. The contact resistance increment estimate is output along with the alarm.

7. The method according to claim 1, characterized in that, During the monitoring phase, dynamically updating the associated baseline at a preset period includes: within a sliding statistical window of the most recent preset number of days, excluding data that has been identified as the transition segment before and after a load switching event and data that has been marked as the branch pairs involved in the suspected branch, and re-statistically calculating the mean and standard deviation using the incremental response ratio of the remaining valid time, replacing the original associated baseline.

8. The method according to claim 1, characterized in that, When the number of branches marked as suspected exceeds a preset proportion of the total number of branches, the coefficient of variation of the correlation deviation score of all branches is calculated. If the coefficient of variation is lower than a preset value, an abnormality prompt is output on the incoming line side. If the coefficient of variation is not lower than the preset value, the individual suspected branches are located according to the correlation deviation score.

9. A multi-branch current and voltage fusion monitoring system for a low-voltage cable branch box, characterized in that, include: The synchronous sampling and calculation module is used to synchronously sample the current and busbar voltage of each branch sharing the same incoming busbar in the branch box, and calculate the voltage increment of the busbar voltage relative to the previous sampling period and the current increment of each branch current relative to the previous sampling period. The effective time screening and ratio calculation module is used to screen the time when the absolute value of the voltage increment is greater than the minimum effective threshold from each sampling time as the effective time, and calculate the incremental response ratio of the branch formed by any two branches at each effective time. The incremental response ratio is the quotient obtained by dividing the current increment of one branch by the current increment of the other branch. The associated baseline establishment module is used to calculate the mean and standard deviation of the incremental response ratio of each branch pair at all effective times during the baseline learning period, and use the mean and standard deviation as the associated baseline of the branch pair. The suspected branch identification module is used to calculate the standardized deviation of each branch pair during the monitoring phase. For each branch, the absolute values ​​of the standardized deviations of all branch pairs containing that branch are summed to obtain the association deviation score of that branch. Branches whose association deviation scores exceed a preset multiple of the median of the association deviation scores of all branches are marked as suspected branches. The deviation cause determination and alarm module is used to perform first-order differencing on the incremental response ratio time series of the suspected branch within the backtracking window. The deviation cause is determined based on whether there is a step feature in the difference sequence. If a step feature is present, the suspected branch is removed and the associated baseline is updated. If no step feature is present, the branch is confirmed to have poor contact and an alarm is output. Determining whether there is a step feature in the difference sequence includes: taking the point with the largest absolute value in the difference sequence. If the absolute value of this point exceeds a first preset multiple of the standard deviation of the difference sequence, and the absolute values ​​of the difference mean within each preset window before and after this point are lower than a second preset multiple of the standard deviation of the difference sequence, then a step feature is determined to exist, wherein the first preset multiple is greater than the second preset multiple.

Images