Intelligent monitoring and early warning method for safe operation of subway door

Abstract

The invention belongs to the technical field of metro door monitoring and early warning, and particularly discloses an intelligent monitoring and early warning method for safe operation of a metro door, which comprises the following steps of: dividing load states into low, medium and high levels by utilizing metro historical load records and using a statistical quantile method based on historical load distribution, and associating differentiated early warning actions for each load level at the same time. Furthermore, the load grade is determined according to the current load state in the actual operation of the subway, and meanwhile, abnormality judgment is performed on the operation parameters and the limit threshold matched with the corresponding load grade, so that the early warning action associated with the corresponding load grade is triggered, and the pertinence and flexibility of operation early warning of the subway door are realized; therefore, the early warning operation is more adaptive to the running state of the subway door, the problems of false alarm and missing alarm caused by a fixed threshold value can be effectively reduced, and the sensitivity and stability of the early warning operation are ensured.

Description

Technical Field

[0001] The present invention belongs to the technical field of subway door monitoring and early warning, and specifically discloses an intelligent monitoring and early warning method for safe operation of subway doors. Background Art

[0002] As a key safety component of rail transit systems, subway doors' operational status directly impacts passenger safety. If an operational anomaly is not promptly addressed, it can prevent doors from opening or closing properly, impacting smooth boarding and alighting. In extreme cases, it can even cause doors to become stuck or accidentally open, threatening passenger lives. Therefore, real-time monitoring and early warning of subway doors are essential during subway operations.

[0003] Solutions have also been proposed in the prior art. For example, Chinese patent publication number CN107219837A discloses an intelligent system for monitoring and warning subway screen door faults. When a subway screen door component or element reaches a warning setting or a subway screen door fails, the platform operation monitoring module or the door control system monitoring module transmits the issued warning information or fault information to the fault diagnosis module and the neural network fault warning module for classification and rating, and then feeds the results back to the user through a display; the fault type and fault address can be determined, and early warning of foreseeable faults can be issued.

[0004] Current subway door monitoring and early warning systems primarily rely on fixed thresholds for anomaly detection. This approach fails to fully account for the dynamic characteristics of door opening and closing under varying load conditions, impacting the reliability and efficiency of early warnings. Specifically, during peak hours, subway doors frequently open and close, significantly increasing the probability of anomalies. Using fixed thresholds for anomaly detection may result in an increased risk of missed detections due to the set thresholds failing to adapt to the high frequency of operation. For example, during peak hours, when doors frequently open and close, fixed thresholds cannot effectively identify potential faults caused by mechanical fatigue, increasing the likelihood of missed detections.

[0005] During off-peak hours, subway doors open and close less frequently, making anomalies less likely. However, because fixed thresholds aren't adjusted to the normal operating range under low-load conditions, this can lead to an increased rate of false alarms. For example, minor parameter fluctuations (such as a slight increase in motor temperature) can trigger an alarm under low-load conditions, even though such fluctuations are normal under these conditions.

[0006] Furthermore, the current existing systems for subway doors mostly use unified warning actions and are unable to respond differently according to load levels. In particular, occasional abnormalities under low-load conditions may trigger unnecessary emergency shutdowns, which not only affects overall operational efficiency but also wastes human and material resources caused by shutdown maintenance. Summary of the Invention

[0007] To this end, one purpose of an embodiment of the present application is to provide an intelligent monitoring and early warning method for safe operation of subway doors that can dynamically sense load status, adaptively adjust thresholds and achieve multi-level early warnings, so as to improve the reliability and efficiency of subway door operation early warnings, thereby solving the problems mentioned in the background technology.

[0008] The purpose of the present invention can be achieved by the following technical solutions: an intelligent monitoring and early warning method for safe operation of subway doors, comprising the following steps: (1) defining the load status indicator as the number of times the door body is opened and closed in a set unit time and the passenger flow, thereby dividing the load status into three levels: low, medium and high based on the statistical quantile method of historical load distribution, and at the same time associating an early warning action for each load level.

[0009] (2) Differentiated operating parameter limit thresholds are configured for each load level based on the operating parameter fault manifestation characteristics in historical fault records under different load levels. The operating parameter fault manifestation characteristics are the time series data of the operating parameters in the set time window before the fault occurs.

[0010] (3) The door operating parameters are collected in real time from the subway door control terminal interface, and the load level is determined according to the current load status indication. At the same time, the collected operating parameters are matched with the limit threshold of the corresponding load level for abnormal judgment.

[0011] (4) When an abnormal operation is determined, the warning action associated with the load level is triggered.

[0012] (5) Periodically count the fault frequencies of each load level and dynamically update the operating parameter limit threshold configuration accordingly.

[0013] Combining all the above technical solutions, the positive effects of the present invention are as follows: 1. The present invention uses historical load records and adopts a statistical quantile method to divide the load status into three levels: low, medium, and high, and associates specific warning actions with each level. In actual operation, an abnormality judgment is made by comparing the operating parameters with the limit threshold of the corresponding load level, and the corresponding warning action is triggered. This can improve the pertinence and flexibility of the warning, adapt it to the actual operating status, effectively reduce false alarms and missed alarms caused by fixed thresholds, and ensure that the warning operation is both sensitive and stable.

[0014] 2. The present invention defines the load status indicator as the number of times the door is opened and closed in a set unit time and the passenger flow, thereby realizing a multi-dimensional evaluation of the load status, thereby more comprehensively reflecting the actual operating conditions, and providing multi-dimensional data support for load level classification, ensuring that the definition of load levels is more reasonable and accurate.

[0015] 3. After implementing differentiated early warning based on load level classification, the present invention dynamically updates the operating parameter limit threshold configuration by periodically counting the fault frequency of each load level, thereby significantly improving the adaptability and accuracy of the system, ensuring that the threshold setting always fits the actual operating conditions, effectively reducing the risk of false alarms and missed alarms, and enhancing system stability and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The present invention is further described with reference to the accompanying drawings. However, the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative effort.

[0017] Figure 1 This is a diagram of the steps for implementing the method of the present invention.

[0018] Figure 2 This is a schematic diagram of the load states of the three tertile subjects corresponding to low, medium and high levels used in the present invention.

[0019] Figure 3 This is a flowchart for the implementation of determining the typical values of door opening and closing times and passenger flow in each unit time within a single day in the present invention. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0021] See also Figure 1 As shown, the present invention proposes an intelligent monitoring and early warning method for safe operation of subway doors, comprising the following steps: (1) defining the load status indicator as the number of times the door is opened and closed in a set unit time and the passenger flow, dividing the load status into three levels: low, medium and high based on the statistical quantile method of historical load distribution, and associating an early warning action for each load level.

[0022] In the exemplary implementation of the above solution, the set unit time may be specifically quantified as one hour.

[0023] It's important to understand that subway door load refers to the operational demands placed on the door during a specific time period. The reason for integrating early warning mechanisms with load status during subway door operation warnings is that the probability of operational failure is closely related to the door's load status. In particular, under high load conditions, frequent opening and closing operations and the passage of large numbers of passengers can lead to increased wear on mechanical components, elevated motor temperatures, and prolonged system response times, increasing the risk of failure. Therefore, integrating early warning mechanisms with actual load status can more accurately reflect the actual operating conditions of subway doors, ensuring that early warning response measures are more tailored to actual needs and improving the accuracy and reliability of the early warning system.

[0024] It's also important to understand that the focus on subway door load status, using opening and closing times and passenger flow as indicators of door load, is primarily based on the following considerations: From an operational perspective, opening and closing times directly reflect the frequency of subway door operations within a specific time period. Frequent opening and closing operations not only increase wear on mechanical components but can also lead to increased motor temperatures and prolonged system response times. Therefore, opening and closing times are a key indicator of subway door load.

[0025] Passenger flow directly affects boarding and alighting frequency, which in turn determines the number of door openings and closings. For example, during peak hours, when a large number of passengers board and alight, doors open and close frequently, significantly increasing the load.

[0026] In terms of data accessibility: Opening and closing times and passenger flow are data that can be easily and accurately collected using existing sensors. For example, infrared array sensors can count the number of passengers entering and exiting a train car, while subway door control systems often have built-in opening and closing counters that display the number of openings and closings directly on the control terminal interface.

[0027] Combining door opening and closing times and passenger flow allows for a comprehensive, multi-dimensional assessment of subway door load status. While a single metric (e.g., door opening and closing times or passenger flow alone) may not fully reflect actual door load conditions, combining the two provides a more accurate load assessment, providing multi-dimensional data support for load classification, ensuring more reasonable and accurate load classification.

[0028] In a preferred implementation of the above-mentioned solution of the present invention, the load status is classified into three levels: low, medium, and high, based on the statistical quantile method of historical load distribution. The following process is used: Load records for each unit time of each day in a historical year are extracted from the subway operation database. Each load record contains load status indicators. The load status indicators in each historical load record for each year are then integrated into a door opening and closing frequency dataset and a passenger flow dataset. Ensure that each annual dataset contains records for all unit times within a complete daily cycle.

[0029] The aforementioned subway operations database integrates multiple data sources, including door control systems and infrared array sensors, to record door opening and closing times and passenger flow in real time. The door control system provides hourly door opening and closing times, while the infrared array sensors collect hourly passenger flow data. This data is cleaned and pre-processed, then converted into load records and stored in the subway operations database.

[0030] For the door opening and closing frequency dataset and passenger flow dataset for all years, data of the same unit time were extracted and standard deviation was calculated. Based on this, the typical value of door opening and closing frequency and passenger flow for each unit time in a single day were determined.

[0031] It's important to note that while the length of each day is fixed, passenger demand fluctuates throughout the day, leading to corresponding changes in subway door operating requirements. To capture the daily load variations at different time units within a single day, historical load records for each time unit of time are required to ensure data reliability and representativeness, avoiding the bias and inaccuracy associated with single-date data. By extracting multiple years of daily load records for each time unit of time and calculating the standard deviation of door opening and closing times and passenger flow at each time unit of time across all years, fluctuations can be captured and typical values determined. This method accurately reflects the load variation patterns at different time units within a single day, providing a solid data foundation for subsequent load classification.

[0032] In a further preferred implementation of the above scheme, see Figure 3 As shown, the typical value of the number of door openings and closings and the typical value of passenger flow in each unit time in a single day are determined as follows: the standard deviation of the number of door openings and closings and the standard deviation of passenger flow in each unit time in a single day are compared with the set critical values, where the standard deviation critical value is pre-set and is used to determine whether the data fluctuation is within an acceptable range.

[0033] If the standard deviation of the door opening and closing times and the passenger flow rate in a certain unit time do not exceed the critical value, it indicates that the data fluctuation of the unit time is within the acceptable range, indicating that the operation demand is stable and there is no significant abnormal fluctuation. The mean of the door opening and closing times and passenger flow rate corresponding to all annual data sets of the unit time is calculated as the typical value of the unit time. Otherwise, it indicates that the data fluctuation of the unit time is large and the use of the mean as the typical value is not representative. In this case, the trimmed mean of the data sets corresponding to all years of the unit time is calculated as the typical value of the unit time.

[0034] It's important to note that the trimmed mean, sometimes also called the truncated mean, mentioned above, is a robust statistic used to reduce the impact of outliers on the mean calculation. It does this by removing a certain proportion of the maximum and minimum values from the dataset and then calculating the mean of the remaining data. Compared to using the median, using the trimmed mean still utilizes most of the information in the original data, rather than just one or two values in the middle, thus providing a more robust and representative estimate of the central tendency.

[0035] In a specific example, assuming that there are 100 data points in the data set and the removal ratio is 5%, all the values in the data set are first sorted in ascending or descending order, and 5% of the extreme values are removed, that is, the lowest 5 values and the highest 5 values are removed, for a total of 10 data points. After removing the extreme values, the arithmetic mean is calculated using the remaining data points.

[0036] The typical values of door opening and closing times and passenger flow at each unit time in a single day are arranged in ascending order and analyzed using the tertile analysis method to determine the first tertile, second tertile, and third tertile of door opening and closing times and passenger flow, respectively.

[0037] The load levels are classified according to the tertile values calculated above as follows: Low load level: defined as the number of door openings and closing times and passenger flow both being lower than their respective first tertiles or one of them being lower than the first tertile and the other being between its first and third tertiles, representing the load level in the low direction.

[0038] Medium load level: defined as the number of door openings and closings and passenger flow rates both being between the first and third tertiles, respectively, indicating the middle-pointing third of the load level.

[0039] High load level: defined as the number of door openings and closings or passenger flow exceeding the respective third tertiles, or one of them being lower than the first tertile and the other higher than the third tertile, representing a high-directed third load level.

[0040] It should be noted that the tertile analysis method is a statistical method that divides a data set into three equal parts. By determining the first tertile, the second tertile (i.e., the median), and the third tertile, the data distribution can be effectively divided.

[0041] In the specific implementation of the above operation, the implementation steps of the tertile analysis method are as follows: the number of switches and passenger flow in each unit time are regarded as independent data points and arranged in ascending order.

[0042] Calculate the quantile position: the first tertile corresponds to the 25% position of the arrangement order.

[0043] The second tertile corresponds to the 50th percentile position in the ranking order.

[0044] The third tertile corresponds to the 75th percentile position of the ranking order.

[0045] In an example, suppose the number of door openings and closing times in 24 time units in a day is arranged in ascending order as follows: [10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240]. The first tertile corresponds to the 25% position of the arrangement order, which is the 6th value, approximately 60 times.

[0046] The second tertile corresponds to the 50% position of the arrangement order, which is the 12th value, approximately 120 times.

[0047] The third tertile corresponds to the 75% position of the arrangement order, which is the 18th value, approximately 180 times.

[0048] See also Figure 2 As shown, the first tertile corresponds to a low load level, the second tertile corresponds to a medium load level, and the third tertile corresponds to a high load level. However, the division based on the subject may lead to data omission or unclear classification problems.

[0049] To this end, the present invention optimizes the specific division of load levels based on the main division: the low load level is suitable for situations where the pressure in at least one dimension is low, and the pressure in a certain dimension is allowed to be higher but still classified as a low load level, thereby avoiding unnecessary warnings or resource allocations due to temporary peaks in a single dimension. Although this situation rarely occurs, it helps to improve the response efficiency of the system.

[0050] The medium load level is suitable for situations where the pressure in both dimensions is at a medium level. It can accurately identify the load status within the normal operating range and is suitable for most daily operation scenarios.

[0051] The high load level applies to situations where the pressure in at least one dimension is high or extremely unbalanced (one dimension is very low and another dimension is very high). This definition can effectively identify high load conditions, especially in cases of extreme imbalance. Although this situation is relatively rare, timely identification and handling of these high-risk situations is critical to ensuring the safety and stability of the system.

[0052] The above-mentioned optimization division method based on tertiles ensures that each data point can be clearly classified into one of the three load levels of low, medium, and high, avoiding data omissions or unclear classification.

[0053] The above method not only comprehensively covers all data points, but also provides detailed and reasonable classification according to the stress levels in different dimensions, improving the accuracy of the early warning mechanism and optimizing resource allocation.

[0054] As an operational example of the above scheme, the tertiles of door opening and closing times and passenger flow are shown in Table 1.

[0055] Table 1: Tertiles of door opening and closing times and passenger flow

[0056]

[0057] For one data point (door opening and closing times are 28, and passenger flow is 300): the door opening and closing times are less than the first tertile, and the passenger flow is between the first tertile and the second tertile.

[0058] According to the above load level classification rules, this data point belongs to the low load level.

[0059] For another data point (door opening and closing times are 47, and passenger flow is 560): the door opening and closing times are between the second and third tertiles, and the passenger flow is greater than the third tertile.

[0060] According to the above load level classification rules, this data point belongs to the high load level.

[0061] In the preferred embodiment of the present invention, each load level is associated with the following warning actions: a low load level is associated with a blue warning, a blue static icon is displayed on the control terminal interface, the current operating parameters are recorded in a log file, and an operation and maintenance work order is not triggered.

[0062] It is important to understand that at low load levels, the fluctuation range of subway door operating parameters (such as resistance and temperature) is relatively small, but occasional anomalies may occur due to transient interference (such as obstruction by foreign objects). For example, if the low-load resistance threshold is 15N, a single instantaneous value of 16N (6.7% deviation) may be caused by sensor noise rather than a real fault. If this situation triggers an operation and maintenance work order, it will lead to unnecessary waste of resources, including manpower scheduling, equipment inspection and downtime losses, and the benefits of fault repair at low loads may be lower than the costs. By recording abnormal data in logs at low load levels for periodic analysis, the frequent generation of invalid work orders can be avoided. Operation and maintenance personnel can regularly review logs to identify occasional anomalies and decide whether further inspection is required, thereby optimizing resource utilization and improving maintenance efficiency.

[0063] The medium load level is associated with a yellow warning, and a yellow flashing icon is dynamically displayed on the control terminal interface. Repeated pop-up windows display abnormal operating parameter information, which can be the abnormal operating parameter name, real-time value and deviation percentage. A pre-maintenance work order is automatically generated and pushed to the operation and maintenance system.

[0064] It's important to understand that a medium load level typically represents a state where subway doors are experiencing moderate operational demands but a certain amount of pressure. Abnormal operating parameters at this point may indicate an increased risk of potential failure. Compared to low load levels, abnormalities under medium load are more likely to be related to actual operating conditions (such as mechanical fatigue or temperature rise caused by frequent opening and closing), rather than simply caused by transient interference or sensor noise. Therefore, by dynamically displaying a yellow warning icon, displaying abnormal parameter information (such as name, real-time value, and deviation percentage) in a pop-up window, and automatically generating a pre-maintenance work order, operations and maintenance personnel can be promptly alerted to intervene in potential issues. This approach allows preventive measures to be taken before a failure occurs, reducing the likelihood of a failure escalating into a serious problem.

[0065] The high load level is associated with a red warning, which covers the entire screen of the control terminal interface with a red warning interface, displays the emergency fault logo and fault location information, and forcibly locks the door to a safe state.

[0066] It's important to understand that high loads typically indicate frequent subway door operation or extreme pressure. Abnormal operating parameters can pose a direct threat to passenger safety and system stability. Abnormalities under high load (such as a sharp increase in resistance or excessive temperatures) are often closely related to mechanical fatigue, component wear, or sudden failure, and carry a high risk of escalating into serious accidents. Therefore, by providing a full-screen warning and forcibly locking the door to a safe state, potentially dangerous behavior can be immediately prevented, ensuring the safety of passengers and equipment. Furthermore, clear fault location information facilitates rapid response and precise repair, minimizing the impact of the fault on operations and ensuring the efficiency and safety of the subway system.

[0067] In a further preferred embodiment of the present invention, (2) differentiated operating parameter limit thresholds are configured for each load level based on operating parameter fault manifestation characteristics in historical fault records at different load levels, wherein the operating parameter fault manifestation characteristics are time series data of the operating parameters within a set time window before the fault occurs. For example, the set time window may be 5 minutes.

[0068] The specific configuration process is as follows: collect historical fault records of subway doors and their associated load status indicators, and classify each historical fault record into the corresponding load level according to the load level classification method.

[0069] The above-mentioned association of historical fault records with load status indicators is achieved by aligning the timestamp of the fault occurrence with the timestamp of the load record, thereby searching for data with the same timestamp in the corresponding load record, and extracting the load status indicator at that moment for association.

[0070] All historical fault records are classified according to their corresponding load levels to form a historical fault record set corresponding to each load level.

[0071] The inducement parameters of each fault record are extracted from the historical fault record set corresponding to each load level, and the fault records with the same inducement parameters are classified to form a fault record set corresponding to each inducement parameter under each load level.

[0072] It's important to note that the aforementioned causal parameters refer to key operating parameters that can cause subway door failures. For example, subway door operating parameters include mechanical resistance and motor temperature. When a subway door fails to close due to excessive resistance, mechanical resistance is considered the causal parameter.

[0073] The fault manifestation characteristics of the inducement parameters of each fault record are extracted from the fault record set corresponding to each inducement parameter under each load level, and the fluctuation rate of the inducement parameters in the adjacent time of the fault is calculated based on this. The calculation of the fluctuation rate can be achieved by quantifying the ratio of the change amplitude of the inducement parameters in the adjacent time before and after the fault to the time interval, and then the inducement parameter value at the previous time in the adjacent time corresponding to the maximum fluctuation rate is taken as the limiting threshold of the inducement parameter in the fault record.

[0074] The present invention uses the fluctuation rate to determine the limit threshold, which can effectively identify the sharp changes that occur in a short time before the failure occurs. These changes are often early signals of impending failure. This method can better reflect the trend of dynamic changes than simple threshold setting.

[0075] The reason we further select the inducing parameter value at the preceding time among the adjacent time periods corresponding to the maximum volatility as the limiting threshold for the inducing parameter in the fault record is that the maximum volatility typically occurs when a fault is about to occur or has just occurred. At this time, the inducing parameter value is often close to the system's tolerance limit. By selecting the inducing parameter value at the preceding time among the adjacent time periods corresponding to the maximum volatility, we can capture the critical point at which the system transitions from normal operation to failure. This selection helps set more precise warning thresholds, allowing the system to issue timely alerts when approaching critical states, thereby preventing failures or reducing the losses caused by failures.

[0076] The limit thresholds of the inducement parameters corresponding to each fault record under each load level are compared, and the limit threshold with the highest occurrence frequency is selected as the limit threshold of the inducement parameters under the corresponding load level.

[0077] The statistical analysis above selects the most frequently occurring thresholds, ensuring that the selected values are the most common critical points, reducing the likelihood of false positives and missed negatives. This statistical method helps improve the stability and reliability of the early warning system.

[0078] The above method relies on historical fault data when configuring differentiated operating parameter limit thresholds for each load level, ensuring that the limit thresholds are set based on actual fault conditions rather than theoretical assumptions. This approach makes the limit thresholds more closely aligned with actual operating conditions, significantly improving the accuracy and reliability of the early warning system. In addition, by configuring limit thresholds based on historical fault data, risks can be identified before potential faults occur, and preventive measures such as adjusting operating parameters or performing emergency repairs can be taken, effectively reducing the probability of faults and improving the overall stability and safety of the system. This data-driven approach not only optimizes resource utilization but also enhances the system's ability to cope with complex operating conditions.

[0079] (3) The door operating parameters are collected in real time from the subway door control terminal interface, and the load level is determined according to the current load status indication. At the same time, the collected operating parameters are matched with the limit threshold of the corresponding load level for abnormal judgment.

[0080] It should be pointed out that the control terminal of the subway door can be integrated into the control box, and the operating parameters of the door body will be displayed on the panel interface of the control box, including but not limited to mechanical resistance, motor temperature, motor current, etc.

[0081] In the manner in which the above-mentioned solution can be implemented, the abnormality determination process is as follows: the collected operating parameters are compared with the limit thresholds matching the corresponding load levels. If the actual value of a certain operating parameter reaches or exceeds its limit threshold, the excess amplitude is calculated and the excess duration is recorded.

[0082] In the above implementation example, the calculation of the limit deviation can be achieved by quantifying the deviation between the actual value of the operating parameter and the corresponding limit threshold. This difference reflects the extent to which the operating parameter exceeds the set safety range and is used to evaluate the severity of the limit deviation.

[0083] Based on the monitoring results of the over-limit amplitude and over-limit duration, determine whether the operation is abnormal according to the following conditions: a) The over-limit amplitude reaches or exceeds the preset warning amplitude.

[0084] b) The duration of the overrun reaches or exceeds the preset warning duration.

[0085] When any of the above conditions is met, the operation is judged to be abnormal.

[0086] The above introduces two quantitative indicators: the over-limit amplitude and the over-limit duration. The over-limit amplitude reflects the severity of the parameter deviation from the safe range, while the over-limit duration reflects the persistence of the deviation. Combining these two dimensions for comprehensive judgment can effectively avoid misjudgment caused by instantaneous fluctuations or short-term over-limits compared to the single threshold trigger method, improve the scientificity and accuracy of abnormal judgment, and ensure that the system triggers abnormal alarms only when intervention is really needed.

[0087] In an improved implementation of the above-mentioned implementation method, the abnormality judgment of matching the collected operating parameters with the limit threshold of the corresponding load level also includes the following content: when the operating parameters are judged to be abnormal, the number of all operating parameters judged to be abnormal is counted within a set time window (such as 5 minutes).

[0088] Use relevant algorithms (such as Pearson correlation coefficient, Spearman rank correlation coefficient, etc.) to evaluate whether there is a significant correlation between these abnormal operating parameters. Specifically, a correlation threshold (such as a correlation coefficient greater than 0.7) can be set to determine whether there is a significant correlation.

[0089] If there is a significant correlation, these abnormal operating parameters will be identified as correlated, and the identification information will be output synchronously when the early warning action is triggered.

[0090] It's important to understand that when multiple operating parameters exhibit abnormalities within a short period of time, correlation analysis can reveal the inherent connections and mutual influences between these abnormal parameters. Synchronously outputting correlation indicators during early warning operations helps maintenance personnel quickly identify the root cause of a fault, not just the symptoms. For example, if mechanical resistance and motor current are both abnormal and significantly correlated, maintenance personnel can prioritize checking the status of the motor and drive system, thereby developing a more targeted preventive maintenance plan, rationally allocating maintenance resources, avoiding unnecessary comprehensive inspections, and saving time and costs.

[0091] (4) When an abnormal operation is determined, the warning action associated with the load level is triggered.

[0092] In the innovative implementation of the above scheme, when an abnormal operation is determined, the warning action associated with the load level is triggered and also includes the following content: after the warning action is triggered, the monitoring screen of the relevant door body is retrieved in real time and uploaded to the train control center simultaneously.

[0093] The monitoring images provided by these operations provide first-hand visual information, enabling the control center to respond more quickly. For example, in an emergency, a decision can be made immediately on whether an emergency stop or other emergency measures are needed to ensure passenger safety.

[0094] (5) Periodically count the fault frequencies of each load level and dynamically update the operating parameter limit threshold configuration accordingly.

[0095] As an example of the above implementation, the periodicity may be weekly or monthly.

[0096] In a preferred embodiment of the present invention, the dynamic update of the operating parameter limit threshold configuration is implemented as follows: the number of historical fault records of each load level is collected in each set period to count the historical fault frequency of each load level.

[0097] Compare the fault frequencies for each load level in the current cycle with the historical fault frequencies in previous cycles to identify whether the fault frequencies have increased.

[0098] In a specific implementation of the above solution, identifying whether the fault frequency has increased includes the following: performing mean and standard deviation statistics on the historical fault frequency of each load level in the previous cycle.

[0099] Normalize the fault frequency of each load level in the current cycle with the mean fault frequency of the corresponding load level in the historical cycle, and calculate the normalized score. The expression of the normalized score is: , where Represents a standardized score, which indicates the degree of deviation of the failure frequency of the current cycle relative to the historical data. Indicates the fault frequency of the current cycle, 、 Represent the mean and standard deviation of the failure frequency in the historical period respectively.

[0100] The standardized score is compared with the preset significance threshold. When the standardized score is greater than the significance threshold, it means that the failure frequency of the current cycle is significantly greater than the average failure frequency of the historical cycles, indicating that there may be new risks or problems. In this case, the failure frequency is identified to have increased. Otherwise, it is considered that the failure frequency has not increased significantly.

[0101] It should be noted that when setting the significance threshold, it can be determined based on actual needs and risk tolerance. For example, at high load levels, a lower significance threshold may need to be set to detect potential problems earlier; at low load levels, the threshold can be relaxed to reduce unnecessary alerts.

[0102] This method calculates the mean and standard deviation of historical cycle failure frequencies and normalizes the current cycle's failure frequency against historical data to identify whether the failure frequency is increasing. This approach leverages fundamental statistical principles, making identification more scientific and accurate. The standardized score quantifies the degree of deviation of the current cycle's failure frequency from historical data, providing an objective metric. Compared to relying solely on experience or intuitive judgment, the standardized score avoids subjective bias caused by human factors, ensuring the fairness and reliability of the identification results.

[0103] As a data example for identifying whether the failure frequency is increasing, assume that the failure frequencies of a subway door in the past and current cycles at various load levels are shown in Table 2.

[0104] Table 2: Fault frequency data at various load levels

[0105]

[0106] The normalized score calculated at low load levels was 1.33 < 2, indicating that the failure frequency at low load levels did not increase significantly.

[0107] The normalized score calculated at the medium load level was 1 < 1.5, indicating that the failure frequency at the medium load level did not increase significantly.

[0108] The normalized score calculated at high load levels was 1.67 > 1, indicating that the frequency of failures increased at high load levels.

[0109] If the fault frequency of a certain load level is not increased, it means that the fault frequency is decreasing or remains stable, then the limit threshold of the current operating parameter is maintained. Otherwise, the fault records of the corresponding load level in the current cycle are collected, and then the operating parameter fault manifestation characteristics in the fault records are used to recalculate and update the operating parameter limit threshold.

[0110] Given the dynamic operating state of subway door systems and the wear and aging that can occur over time, their performance can change accordingly. By periodically assessing failure frequencies and adjusting operating parameter limit thresholds, we ensure that the system adapts to these changes, maintaining efficient and stable early warning operations, thereby ensuring system reliability and safety.

[0111] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0112] The above embodiments may be implemented in whole or in part through software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0113] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0114] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0115] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0116] Finally, 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 spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An intelligent monitoring and early warning method for safe operation of subway doors, characterized in that: The following steps are involved: (1) Define the load status indicator as the number of times the door is opened and closed in a set unit time and the passenger flow. Based on the statistical quantile method of historical load distribution, the load status is divided into three levels: low, medium, and high. At the same time, an early warning action is associated with each load level. (2) configuring differentiated operating parameter limit thresholds for each load level based on the operating parameter fault manifestation characteristics in historical fault records under different load levels, wherein the operating parameter fault manifestation characteristics are time series data of the operating parameters in a set time window before the fault occurs; (3) Collect door operating parameters in real time from the subway door control terminal interface, determine the load level based on the current load status indication, and perform abnormality judgment by matching the collected operating parameters with the limit threshold of the corresponding load level; (4) When an abnormal operation is determined, a warning action associated with the load level is triggered; (5) Periodically count the fault frequencies of each load level and dynamically update the operating parameter limit threshold configuration accordingly.

2. The intelligent monitoring and early warning method for safe operation of subway doors according to claim 1, characterized in that: The statistical quantile method based on historical load distribution divides the load status into three levels: low, medium, and high. See the following process: Extract the load records of each unit time per day in each historical year from the subway operation database, and integrate the load status indicators in the historical load records of each year into a door opening and closing times dataset and a passenger flow dataset respectively; For all annual door opening and closing frequency datasets and passenger flow datasets, we extracted data of the same unit time and calculated the standard deviation. Based on this, we determined the typical values of door opening and closing frequency and passenger flow for each unit time within a single day. The typical values of door opening and closing times and passenger flow at each unit time in a single day are arranged in ascending order and analyzed using the tertile analysis method to determine the first tertile, second tertile, and third tertile of door opening and closing times and passenger flow respectively; The load levels are classified according to the tertile values calculated above as follows: Low load level: defined as the door opening and closing times and passenger flow are both lower than the first tertile, or one of them is lower than the first tertile and the other is between the first and third tertiles, representing the low third load level; Medium load level: defined as the door opening and closing times and passenger flow are between the first and third tertiles, indicating the middle point to the third load level; High load level: defined as the number of door openings and closings or passenger flow exceeding the respective third tertiles, or one of them being lower than the first tertile and the other higher than the third tertile, representing a high-directed third load level.

3. The intelligent monitoring and early warning method for safe operation of subway doors according to claim 2, characterized in that: The process of determining the typical value of door opening and closing times and passenger flow rate in each unit time in a single day is as follows: Compare the standard deviation of door opening and closing times and passenger flow rate at each unit time in a single day with the set critical values; If the standard deviation of the door opening and closing times and the standard deviation of the passenger flow in a unit time do not exceed the critical value, the mean of the door opening and closing times and the passenger flow in all annual data sets corresponding to the unit time are calculated as the typical value of the unit time. Otherwise, the trimmed mean of the data sets corresponding to all annual data sets of the unit time are calculated as the typical value of the unit time.

4. The intelligent monitoring and early warning method for safe operation of subway doors according to claim 1, characterized in that: The following warning actions are associated with each load level: A low load level is associated with a blue warning, a blue static icon is displayed on the control terminal interface, and the current operating parameters are recorded in a log file without triggering an operation and maintenance work order; A yellow warning is associated with the medium load level, and a yellow flashing icon is dynamically displayed on the control terminal interface. A pop-up window displays abnormal operating parameter information repeatedly, and a pre-maintenance work order is automatically generated and pushed to the operation and maintenance system. The high load level is associated with a red warning, which covers the entire screen of the control terminal interface with a red warning interface, displays the emergency fault logo and fault location information, and forcibly locks the door to a safe state.

5. The intelligent monitoring and early warning method for safe operation of subway doors according to claim 1, characterized in that: The process of configuring differentiated operating parameter limit thresholds for each load level is as follows: Collect historical subway door fault records and their associated load status indicators, and classify each historical fault record into the corresponding load level according to the load level classification method; Classify all historical fault records according to their corresponding load levels to form a historical fault record set corresponding to each load level; Extracting the cause parameters of each fault record from the historical fault record set corresponding to each load level, and classifying the fault records with the same cause parameters to form a fault record set corresponding to each cause parameter under each load level; Extract the fault manifestation characteristics of the inducing parameters of each fault record from the fault record set corresponding to each inducing parameter at each load level, and calculate the fluctuation rate of the inducing parameters in the adjacent time of the fault based on this. Then, take the inducing parameter value at the previous time in the adjacent time corresponding to the maximum fluctuation rate as the limiting threshold of the inducing parameter in the fault record; The limiting thresholds of the inducement parameters corresponding to each fault record under each load level are compared, and the limiting threshold with the highest occurrence frequency is selected as the limiting threshold of the inducement parameters under the corresponding load level.

6. The intelligent monitoring and early warning method for safe operation of subway doors according to claim 1, characterized in that: The following process is used to determine abnormalities by combining the collected operating parameters with the limit thresholds matching the corresponding load levels: Compare the collected operating parameters with the limit thresholds matching the corresponding load levels. If the actual value of a certain operating parameter reaches or exceeds its limit threshold, calculate the excess magnitude and record the duration of the excess. Based on the monitoring results of the over-limit magnitude and over-limit duration, determine whether there is an operation abnormality according to the following conditions: a) The limit-exceeding range reaches or exceeds the preset warning range; b) The duration of the overrun reaches or exceeds the preset warning duration; When any of the above conditions is met, the operation is judged to be abnormal.

7. The intelligent monitoring and early warning method for safe operation of subway doors according to claim 6, characterized in that: The abnormality determination based on the collected operating parameters and the limit thresholds matched with the corresponding load levels also includes the following: When an operating parameter is judged to be abnormal, the number of all operating parameters judged to be abnormal is counted within the set time window; The relevant algorithm is used to evaluate whether there is a significant correlation between abnormal operating parameters. If there is a significant correlation, these abnormal operating parameters are identified as correlated, and the identification information is output synchronously when the early warning action is triggered.

8. The intelligent monitoring and early warning method for safe operation of subway doors according to claim 1, characterized in that: The triggering of the warning action associated with the load level when determining abnormal operation also includes the following: After the early warning action is triggered, the monitoring screen of the relevant door body is retrieved in real time and uploaded to the train control center simultaneously.

9. The intelligent monitoring and early warning method for safe operation of subway doors according to claim 5, characterized in that: The dynamic update operation parameter limit threshold configuration is implemented as follows: Collect the number of historical fault records of each load level in each set period and calculate the historical fault frequency of each load level; Compare the fault frequency of each load level in the current cycle with the historical fault frequency of the previous cycle to identify whether the fault frequency has increased; If the fault frequency of a certain load level is not increased, the limit threshold of the current operating parameter is maintained. Otherwise, the fault records of the corresponding load level in the current cycle are collected, and then the operating parameter fault characteristics in the fault records are used to recalculate and update the operating parameter limit threshold.

10. The intelligent monitoring and early warning method for safe operation of subway doors according to claim 9, characterized in that: The identification of whether the fault frequency increases includes the following: Compute the mean and standard deviation of the historical failure frequency of each load level in the previous cycle; Normalize the fault frequency of each load level in the current cycle with the mean fault frequency of the corresponding load level in the historical cycle to calculate the normalized score; The standardized score is compared with the preset significance threshold. When the standardized score is greater than the significance threshold, it is considered that the fault frequency has increased. Otherwise, it is considered that the fault frequency has not increased significantly.

Images