A fault detection method, apparatus, electronic device, and storage medium

Abstract

This invention discloses a fault detection method, apparatus, electronic device, and storage medium, relating to the field of data processing technology. The fault detection method includes: acquiring control operation data and auxiliary operation data of a vehicle door system; determining the control statistical characteristics of the control operation data and the auxiliary statistical characteristics of the auxiliary operation data; acquiring historical healthy operation data and historical healthy auxiliary data as health reference sample data; determining the health statistical characteristics of the health reference sample data; determining target clusters of the health statistical characteristics using a target clustering algorithm; determining target detection threshold intervals for the control statistical characteristics and auxiliary statistical characteristics based on the cluster feature information of the target clusters; and determining the vehicle door health status and vehicle door fault mode based on the control statistical characteristics, auxiliary statistical characteristics, and target detection threshold intervals, thereby achieving accurate determination of the vehicle door health status and vehicle door fault mode and improving the efficiency of determining the vehicle door health status and vehicle door fault mode.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a fault detection method, apparatus, electronic device, and storage medium. Background Technology

[0002] With the continuous expansion of urban rail transit networks, train door systems, as a key interface for passengers to enter and exit trains, directly affect operational safety and service quality. Existing door systems typically consist of electric drive mechanisms, control units, guide rail assemblies, and limit devices, and are frequently opened and closed during train arrival and departure.

[0003] In existing train door systems, the methods for identifying equipment operating status are typically quite simplistic, relying mainly on fixed thresholds set by experts based on experience during the research and development or testing phases. This method usually selects a key performance parameter as the criterion, such as the peak current during motor start-up and closing, the time required for door opening and closing, and the response of limit signals. During system operation, the real-time value of this parameter is collected and compared with preset upper and lower thresholds. When the detected value exceeds the set range, the system determines that the door has an operational abnormality or that the equipment is in an abnormal risk state, requiring maintenance or manual re-inspection. Due to factors such as mechanical wear, structural aging, and environmental disturbances from long-term operation, door systems are prone to structural degradation problems such as door leaf misalignment, guide rail jamming, and unstable drive. Especially in the early stages, abnormal characteristics are often hidden in parameter fluctuations, making timely identification through manual inspection difficult. Therefore, how to accurately identify faults in the door system has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a fault detection method, device, electronic device, and storage medium to solve the problem that manual inspection cannot identify door system faults in a timely manner.

[0005] According to one aspect of the present invention, a fault detection method is provided, wherein the method includes:

[0006] Acquire control operation data and auxiliary operation data of the vehicle door system, and determine the control statistical characteristics of the control operation data and the auxiliary statistical characteristics of the auxiliary operation data;

[0007] Historical health operation data and historical health auxiliary data are obtained as health reference sample data, and the health statistical characteristics of the health reference sample data are determined.

[0008] The target clusters of the health statistical features are determined by the target clustering algorithm, and the target detection threshold ranges of the control statistical features and the auxiliary statistical features are determined based on the cluster feature information of the target clusters.

[0009] The door health status and door failure mode are determined based on the control statistical features, the auxiliary statistical features, and the target detection threshold range.

[0010] According to another aspect of the present invention, a fault detection device is provided, wherein the device comprises:

[0011] The feature acquisition module is used to acquire control operation data and auxiliary operation data of the door system, and determine the control statistical features of the control operation data and the auxiliary statistical features of the auxiliary operation data.

[0012] The feature determination module is used to acquire historical health operation data and historical health auxiliary data as health reference sample data, and to determine the health statistical features of the health reference sample data.

[0013] The interval determination module is used to determine the target cluster of the health statistical features through a target clustering algorithm, and to determine the target detection threshold interval of the control statistical features and the auxiliary statistical features based on the cluster feature information of the target cluster.

[0014] The fault detection module is used to determine the health status and fault mode of the door based on the control statistical features, the auxiliary statistical features and the target detection threshold range.

[0015] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0016] At least one processor; and

[0017] A memory communicatively connected to the at least one processor; wherein,

[0018] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the fault detection method according to any embodiment of the present invention.

[0019] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the fault detection method according to any embodiment of the present invention.

[0020] According to another aspect of the present invention, embodiments of the present invention also provide a computer program product, the computer program product including a computer program, which, when executed by a processor, implements the fault detection method of any embodiment of the present invention.

[0021] The technical solution of this invention acquires control operation data and auxiliary operation data of the vehicle door system, determines the control statistical characteristics of the control operation data and the auxiliary statistical characteristics of the auxiliary operation data, acquires historical health operation data and historical health auxiliary data as health reference sample data, determines the health statistical characteristics of the health reference sample data, determines the target cluster of the health statistical characteristics through a target clustering algorithm, determines the target detection threshold range of the control statistical characteristics and auxiliary statistical characteristics based on the cluster feature information of the target cluster, realizes the dynamic construction of the target detection threshold range, determines the door health status and door fault mode based on the control statistical characteristics, auxiliary statistical characteristics and target detection threshold range, and integrates multiple control operation data and auxiliary operation data for parallel discrimination to improve the perception capability of structural faults, realize the accurate determination of the door health status and door fault mode, and improve the efficiency of determining the door health status and door fault mode.

[0022] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart of a fault detection method provided in Embodiment 1 of the present invention;

[0025] Figure 2 This is a flowchart of a fault detection method provided according to Embodiment 2 of the present invention;

[0026] Figure 3 This is a flowchart of a fault detection method provided in Embodiment 3 of the present invention;

[0027] Figure 4 This is a schematic diagram of the structure of a fault detection device according to Embodiment 4 of the present invention;

[0028] Figure 5 This is a schematic diagram of the structure of an electronic device that implements the fault detection method of the present invention. Detailed Implementation

[0029] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0030] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0031] Example 1

[0032] Figure 1 This is a flowchart of a fault detection method according to Embodiment 1 of the present invention. This embodiment is applicable to fault detection of doors in rail transit systems. The method can be executed by a fault detection device, which can be implemented in hardware and / or software and can be configured in an electronic device. Figure 1 As shown, the method includes:

[0033] S110. Acquire control operation data and auxiliary operation data of the door system, and determine the control statistical characteristics of the control operation data and the auxiliary statistical characteristics of the auxiliary operation data.

[0034] The door system can be understood as the device responsible for opening and closing the doors. Control operation data refers to the raw signal data that directly controls the opening and closing process. Generally, control operation data includes at least motor current, door angle, opening and closing speed, and door operator output torque. Auxiliary operation data refers to auxiliary parameters that do not directly drive the door movement but affect the core operating state. These mainly reflect external or indirect influencing factors such as environment, power supply, and losses. For example, auxiliary operation data may include, but is not limited to, ambient temperature, power supply voltage, door lock signal status, and control command response time. In actual operation, the door system can be divided into multiple operating stages during door opening and closing, such as the starting stage, constant speed stage, and deceleration stage. Control operation data and auxiliary operation data can be acquired separately for each operating stage. Control statistical characteristics can be understood as the statistical characteristics of control operation data. Generally, control statistical characteristics can be the time-domain statistical characteristics of control operation data. Auxiliary statistical characteristics can be understood as the statistical characteristics of auxiliary operation data. Generally, auxiliary statistical characteristics can be the time-domain statistical characteristics of auxiliary operation data.

[0035] In this embodiment, pre-set sensors can collect data such as motor current, door angle, opening / closing speed, and door operator output torque during the opening and closing process of the vehicle door system as control operation data. Ambient temperature, power supply voltage, door lock signal status, and control command response time are also collected as auxiliary operation data. Time-domain statistical features of both control and auxiliary operation data are extracted, with the control operation data's time-domain statistical features used as control statistical features and the auxiliary operation data's time-domain statistical features used as auxiliary statistical features. In actual operation, the opening and closing process of the vehicle door system can be divided into multiple operating stages based on the opening and closing speed. Data such as motor current, door angle, opening / closing speed, and door operator output torque for each operating stage are acquired as control operation data, while ambient temperature, power supply voltage, door lock signal status, and control command response time are collected as auxiliary operation data for each operating stage. Time-domain statistical features for each type of data are determined, such as extreme values, mean values, rate of change, duration, and fluctuation amplitude. For different control operation data and auxiliary operation data, corresponding time-domain features that need to be statistically analyzed can be preset, and the corresponding time-domain features can be calculated as control statistical features and auxiliary statistical features respectively. For example, the motor current can be determined by the starting peak value (extreme value), average current (mean value), time from start to peak value (rate of change), and standard deviation of fluctuation (amplitude of fluctuation); the gate rotation angle can be determined by the maximum rotation angle (amplitude), total opening and closing time (duration), and rate of change of rotation angle in the constant speed segment (rate of change), etc., which can be set according to business needs. In one embodiment, when the control operation data and auxiliary operation data include data from multiple operation stages, time-domain statistical features can be calculated for the data of each operation stage separately.

[0036] S120. Obtain historical health operation data and historical health auxiliary data as health reference sample data, and determine the health statistical characteristics of the health reference sample data.

[0037] Historical health operation data can be understood as control operation data where the door health status is determined to be healthy at a historical time. Historical health auxiliary data can be understood as auxiliary operation data where the door health status is determined to be healthy at a historical time. Both historical health operation data and historical health auxiliary data can be divided into multiple operation stages, including data from multiple operation stages. Health reference sample data refers to sample data used to determine the target detection threshold ranges for control statistical features and auxiliary statistical features. Health reference sample data includes historical health operation data and historical health auxiliary data. Health statistical features can be understood as the statistical features of health reference sample data. Generally, health statistical features can be the time-domain statistical features of health reference sample data. When health reference sample data includes data from multiple operation stages, time-domain statistical features can be calculated for each operation stage separately. Each health statistical feature can contain a set of time-domain features of historical health operation data and historical health auxiliary data at the same time.

[0038] In this embodiment, control operation data and auxiliary operation data showing healthy door health status at historical moments can be extracted. The control operation data showing healthy door health status at historical moments is used as historical healthy operation data, and the auxiliary operation data showing healthy door health status at historical moments is used as historical healthy auxiliary data. These historical healthy operation data and historical healthy auxiliary data are then used as health reference sample data, and the temporal statistical characteristics of the health reference sample data are determined as health statistical features. In actual operation, the dimensions of the health statistical features can be the same as the dimensions of the control statistical features of the control operation data and the auxiliary statistical features of the auxiliary operation data in the above steps, and the determination method is also the same.

[0039] S130. Determine the target clusters of health statistical features through the target clustering algorithm, and determine the target detection threshold range of control statistical features and auxiliary statistical features based on the cluster feature information of the target clusters.

[0040] The target clustering algorithm can be understood as an algorithm used to cluster health statistical features. In practice, the target clustering algorithm uses intra-cluster variance as a key indicator to measure the stability of the cluster structure and introduces it into the clustering objective optimization function for joint optimization. In practical applications, the objective optimization function can consist of two parts: the first part is the compactness objective of traditional K-means clustering, and the second part is the intra-cluster feature variance constraint. This ensures the similarity of samples within the cluster while constraining the fluctuation range of features, ultimately providing an accurate target detection threshold range for door fault diagnosis. A target cluster can be understood as a set that groups similar health statistical features into one class. The number of target clusters can be preset according to business needs. Cluster feature information can be understood as the core information of the features in the target cluster, such as the intra-cluster mean, intra-cluster standard deviation, and the variance of health statistical features in each dimension. The target detection threshold range refers to the threshold range used to determine the health status and fault mode of the door. In practice, the corresponding target detection threshold range can be calculated separately for each dimension of control statistical features and auxiliary statistical features.

[0041] In this embodiment, a target number of initial clusters can be set, and a target number of health statistical features can be randomly selected as initial cluster centers. For each health statistical feature, all initial clusters are iterated through. The squared distance from each health statistical feature to the initial cluster center is determined. The sum of the variances of the health statistical features in each dimension within the current initial cluster is determined as the variance value. The product of the variance value and a preset coefficient is determined as the variance optimization term. The sum of the squared distance and the variance optimization term is used as the cost value, resulting in the cost value from each health statistical feature to each initial cluster. The health statistical features are assigned to the initial cluster with the minimum cost value to form intermediate clusters. The mean of each dimension of each intermediate cluster is then calculated and used as the new target cluster center to update the target cluster center. Based on the health statistical features in the new intermediate clusters, the sum of the variances of each dimension of each cluster is recalculated until the rate of change of the total cost value (or average cost value) of all health statistical features in the current iteration is less than the target value. The intermediate clusters corresponding to each health statistical feature are then determined as the target clusters.

[0042] S140. Determine the door health status and door fault mode based on control statistical characteristics, auxiliary statistical characteristics, and target detection threshold range.

[0043] The door health status can be used to indicate whether the door system is healthy. For example, the door health status can include healthy, sub-healthy, and faulty. The fault mode is used to indicate the faults existing in the door system.

[0044] In this embodiment, the target detection threshold range corresponding to each control statistical feature and auxiliary statistical feature can be determined separately. It can be determined whether the control statistical feature and auxiliary statistical feature of each dimension fall within the corresponding target detection threshold range. If so, the control statistical feature / auxiliary statistical feature is determined to be without abnormality; otherwise, it is determined that the control statistical feature / auxiliary statistical feature is abnormal. At this point, the number of abnormal control statistical features / auxiliary statistical features can be determined, and the door health status is determined according to the number of abnormalities. For example, multiple quantity thresholds can be preset. When the number of abnormalities exceeds the first quantity threshold, the door system's health status is determined to be sub-healthy; when the number of abnormalities exceeds the second quantity threshold, the door system's health status is determined to be faulty; and when the number of abnormalities is 0, the door system's health status is determined to be healthy. Generally, the door fault mode corresponding to each feature combination can be pre-stored. Abnormal control statistical features and auxiliary statistical features are determined, and these abnormal control statistical features and auxiliary statistical features are used as feature combinations. Then, the corresponding fault mode is matched with a pre-set fault database as the door fault mode. For example, when there is an abnormality in the peak current during the starting phase and an abnormality in the average speed during the constant speed phase, the door fault mode can be determined to be a starting system fault.

[0045] In this embodiment of the invention, by acquiring control operation data and auxiliary operation data of the vehicle door system, the control statistical characteristics of the control operation data and the auxiliary statistical characteristics of the auxiliary operation data are determined. Historical health operation data and historical health auxiliary data are acquired as health reference sample data, and the health statistical characteristics of the health reference sample data are determined. Target clustering algorithm is used to determine the target clusters of the health statistical characteristics. Based on the cluster feature information of the target clusters, the target detection threshold range of the control statistical characteristics and auxiliary statistical characteristics is determined, realizing the dynamic construction of the target detection threshold range. Based on the control statistical characteristics, auxiliary statistical characteristics, and target detection threshold range, the vehicle door health status and vehicle door fault mode are determined. Multiple control operation data and auxiliary operation data are combined and judged in parallel to improve the perception capability of structural faults, realize the accurate determination of vehicle door health status and vehicle door fault mode, and improve the efficiency of determining vehicle door health status and vehicle door fault mode.

[0046] This includes determining the door health status and fault mode based on control statistical characteristics, auxiliary statistical characteristics, and target detection threshold ranges, as well as:

[0047] Control operation data and auxiliary operation data that determine the health status of the doors are healthy are used as candidate health operation data, and auxiliary operation data that determine the health status of the doors are healthy are used as candidate health auxiliary data;

[0048] The candidate health operation data and candidate health auxiliary data are used as candidate health reference sample data;

[0049] When the number of candidate health reference sample data exceeds a preset threshold, the candidate health reference sample data will be used as the health reference sample data.

[0050] The candidate health reference sample data can be understood as the health reference sample data to be selected. The preset threshold number can be set according to business needs.

[0051] In this embodiment, control operation data and auxiliary operation data showing a healthy door health status can be extracted as candidate healthy operation data, and auxiliary operation data showing a healthy door health status can be extracted as candidate healthy auxiliary data. These candidate healthy operation data and candidate healthy auxiliary data are then used as candidate healthy reference sample data. When the number of candidate healthy reference sample data exceeds a preset threshold, it can be determined that the target detection threshold range needs to be redefined. In this case, the candidate healthy reference sample data can be used as healthy reference sample data to facilitate the dynamic determination of the target detection threshold range and improve its certainty.

[0052] Example 2

[0053] Figure 2 This is a flowchart of a fault detection method according to Embodiment 2 of the present invention. This embodiment is a further optimization and extension based on the above embodiments, and can be combined with various optional technical solutions in the above embodiments. Figure 2 As shown, the method includes:

[0054] S201. Collect the rotational speed of the door system during the opening and closing process, and divide the opening and closing process into at least one operating stage according to the rotational speed.

[0055] The opening and closing speed can be understood as the change in the speed of the drive motor throughout the entire process of the door system from opening to closing (or closing to opening). The speed range during the opening and closing process can be different for each operating phase. Operating phases can include at least a starting phase, a constant speed phase, and a deceleration phase.

[0056] In this embodiment, the door system can be monitored by sensors to collect the rotational speed during the opening and closing process, and the opening and closing process can be divided into at least one operating stage by the range of the rotational speed during the opening and closing process and the rotational speed of each operating stage.

[0057] S202. Collect the motor current, gate angle, opening and closing speed and gate machine output torque corresponding to each operating stage, and use the motor current, gate angle, opening and closing speed and gate machine output torque of each operating stage as control operation data.

[0058] Among these, motor current is the current flowing through the motor that drives the door system, reflecting the motor's load and operating status. Door rotation angle is the angle the door rotates around the hinge as it moves from fully closed to fully open (or from fully open to fully closed). Opening / closing speed is the rotational speed of the drive motor during the electric opening and closing of the door. Door operator output torque is the torsional torque transmitted from the motor to the door body through the transmission mechanism.

[0059] In this embodiment, the motor current, door angle, opening / closing speed, and door operator output torque corresponding to each operating stage can be determined separately, and these parameters can be used as control operation data for each operating stage. Alternatively, the motor current, door angle, opening / closing speed, and door operator output torque during the door opening and closing process can be collected, and then the corresponding time periods can be broken down according to the operating stage as control operation data.

[0060] S203. Collect the ambient temperature, power supply voltage, door lock signal status, and control command response time for each operating stage, and use the ambient temperature, power supply voltage, door lock signal status, and control command response time for each operating stage as auxiliary operating data.

[0061] The door lock signal status can be used to indicate whether a door lock is present, and can be represented by a Boolean value. The control command response time can be used to indicate the door opening delay time or the door closing delay time.

[0062] In this embodiment, the ambient temperature, power supply voltage, and control command response time of each operating stage can be collected separately to determine the door lock signal status. The ambient temperature, power supply voltage, door lock signal status, and control command response time of each operating stage can be used as auxiliary operating data.

[0063] S204. Extract the time-domain statistical features of the control operation data and the auxiliary operation data respectively, and use the time-domain statistical features of the control operation data as the control statistical features and the time-domain statistical features of the auxiliary operation data as the auxiliary statistical features.

[0064] In the embodiments, the time-domain statistical characteristics of the control operation data and the auxiliary operation data at each operation stage can be determined, such as mean, extreme value, rate of change, duration and fluctuation amplitude, etc. The time-domain statistical characteristics of the control operation data are determined as control statistical characteristics, and the time-domain statistical characteristics of the auxiliary operation data are determined as auxiliary statistical characteristics.

[0065] S205. Extract control operation data and auxiliary operation data of the door health status from historical time points.

[0066] In this embodiment, the control operation data and auxiliary operation data at historical moments can be stored after collection. The control operation data and auxiliary operation data at historical moments that indicate a healthy door health status can be determined, and the control operation data and auxiliary operation data at historical moments that indicate a healthy door health status can be extracted.

[0067] S206. The control operation data that determines the health status of the doors as healthy is used as historical health operation data, and the auxiliary operation data that determines the health status of the doors as healthy is used as historical health auxiliary data. The historical health operation data and the historical health auxiliary data are used as health reference sample data.

[0068] In this embodiment, control operation data showing a healthy door health status can be used as historical health operation data, auxiliary operation data showing a healthy door health status can be used as historical health auxiliary data, and historical health operation data and historical health auxiliary data can be used as health reference sample data.

[0069] S207. Extract the time-domain statistical features of the health reference sample data as health statistical features.

[0070] In this embodiment, the temporal statistical features of each dimension of the health reference sample data can be extracted as health statistical features, and each health statistical feature can contain a set of statistical features of the health reference sample data at the same time.

[0071] S208. Determine at least one target cluster based on health statistical characteristics using a target clustering algorithm.

[0072] In this embodiment, health statistical features can be clustered according to a target clustering algorithm to obtain at least one target cluster.

[0073] Among them, at least one target cluster for determining health statistical characteristics through a target clustering algorithm includes:

[0074] Determine the cluster centers of the initial clusters for the target number of health statistical features, determine the squared distance between each health statistical feature and each cluster center as the distance value, and determine the sum of the variances of each dimension of the health statistical features in the initial clusters in which the health statistical features are located as the variance value.

[0075] The distance and variance values ​​are input into the objective optimization function to determine the cost value, and the initial cluster with the minimum cost value is used as the intermediate cluster for each health statistical feature.

[0076] Determine the mean of the health statistical features in each intermediate cluster, use the mean of the health statistical features as the target cluster center, redetermine the intermediate clusters of each health statistical feature according to the target cluster center, until the rate of change of the cost value is less than the target value, and determine the intermediate clusters corresponding to each health statistical feature as the target clusters.

[0077] The target quantity can be a number set by business personnel according to business needs. The target optimization function can be understood as a function that determines the cost value assigned to each health statistical feature to each cluster. In actual operation, the target optimization function can consist of distance values, variance values, and preset coefficients. In one embodiment, the target optimization function can be... ;in, For value; The target number of initial clusters; This represents the k-th initial cluster; The number of samples in the k-th initial cluster; Health statistical features for a single health reference sample data (including multi-dimensional features such as motor current and door rotation angle). The center (mean) of the kth initial cluster; The squared Euclidean distance from the health statistics feature to the initial cluster center; d represents the preset coefficient; d represents the total number of dimensions of the health statistics features (such as motor current, door rotation angle, opening and closing speed, and door machine output torque, a total of 4 features). Let the j-th feature be the value of the i-th health statistical feature; Let be the mean of the j-th feature of the k-th initial cluster; Let be the variance of the j-th feature of the k-th initial cluster.

[0078] In this embodiment, a target number of health statistical features can be randomly selected as the cluster centers of the initial clusters. The squared distance between each health statistical feature and each cluster center is determined as the distance value, and the sum of the variances of the health statistical features in each dimension within the initial cluster is used as the variance value. The cost value is determined according to the objective optimization function, and the initial cluster with the minimum cost value corresponding to each health statistical feature is selected as the intermediate cluster for that health statistical feature. The mean of the health statistical features in each intermediate cluster is then determined, and this mean is used as the target cluster center. The intermediate clusters for each health statistical feature are then re-determined according to the target cluster centers until the rate of change of the cost value is less than the target value. Finally, the intermediate clusters corresponding to each health statistical feature are selected as the target clusters.

[0079] S209. Determine the mean and standard deviation of the health statistical characteristics of each dimension in each operational stage of the target cluster, and use the mean and standard deviation as cluster feature information.

[0080] In the embodiment, the health statistical features of each dimension in each running stage of each target cluster can be extracted, the mean and standard deviation of the health statistical features of each dimension can be calculated, and the mean and standard deviation can be used as cluster feature information.

[0081] S210. Determine the product of the standard deviation and the preset parameter as the first product, determine the difference between the mean of each cluster feature information and the first product as the lower threshold, and take the sum of the mean of each cluster feature information and the first product as the upper threshold. Combine the lower threshold and the upper threshold to obtain the target detection threshold range.

[0082] The preset parameters refer to the coefficients used to adjust the fluctuation range of the target detection threshold, which can be set according to business needs.

[0083] In the embodiment, the product of the standard deviation and the preset parameter can be determined as the first product, the difference between the mean of each cluster feature information and the first product can be used as the lower threshold, the sum of the mean of each cluster feature information and the first product can be used as the upper threshold, and the lower threshold and the upper threshold can be combined to obtain the target detection threshold range.

[0084] S211. Extract the target detection threshold interval of each control statistical feature as the first detection threshold, and extract the target detection threshold interval corresponding to each auxiliary statistical feature as the second detection threshold.

[0085] In this embodiment, each control statistical feature and auxiliary statistical feature has a corresponding target detection threshold range. The target detection threshold range of each control statistical feature can be used as the first detection threshold, and the target detection threshold range corresponding to each auxiliary statistical feature can be used as the second detection threshold.

[0086] S212. Determine the first attribution relationship between each control statistical feature and the corresponding first detection threshold, and the second attribution relationship between each auxiliary statistical feature and the corresponding second detection threshold, and determine the health status of the car door according to the first attribution relationship and the second attribution relationship.

[0087] In this embodiment, a first attribution relationship between each control statistical feature and its corresponding first detection threshold can be determined, i.e., whether each control statistical feature falls within its corresponding first detection threshold; and a second attribution relationship between each auxiliary statistical feature and its corresponding second detection threshold can be determined, i.e., whether each auxiliary statistical feature falls within its corresponding second detection threshold. When a control statistical feature falls within its corresponding first detection threshold, the attribution relationship is determined to be attribution; otherwise, it is not attribution. When an auxiliary statistical feature falls within its corresponding second detection threshold, the attribution relationship is determined to be attribution; otherwise, it is not attribution. The number of non-attribution relationships in the first and second attribution relationships is determined, and the door health status is determined based on this number. For example, when the number of non-attribution relationships is 0, the door health status can be determined to be healthy; when the number of non-attribution relationships is greater than a preset first threshold and less than or equal to a preset second threshold, the door health status can be determined to be sub-healthy; and when the number of non-attribution relationships is greater than the preset second threshold, the door health status can be determined to be faulty. The preset first and preset second thresholds can be set according to business requirements.

[0088] S213. Determine the first attribution relationship as non-attribution control statistical feature as the first control statistical feature, determine the second attribution relationship as non-attribution auxiliary statistical feature as the second control statistical feature, and take the combination of the first control statistical feature and the second control statistical feature as the feature combination.

[0089] In an embodiment, a control statistical feature with a first attribution relationship of "not attribution" can be determined as a first control statistical feature, an auxiliary statistical feature with a second attribution relationship of "not attribution" can be determined as a second control statistical feature, and the combination of all the first and second control statistical features can be used as a feature combination.

[0090] S214. Match the fault patterns in the preset fault database according to the feature combination, and take the successfully matched fault patterns as the door fault patterns.

[0091] The preset fault library can be pre-stored faults, and each fault can have a mapping relationship with a combination of features.

[0092] In this embodiment, a fault mode that matches the feature combination can be determined from a preset fault database, and the successfully matched fault mode can be used as the door fault mode.

[0093] In this embodiment of the invention, the opening and closing speed of the vehicle door system is collected during the opening and closing process. The opening and closing process is divided into at least one operating stage according to the opening and closing speed. The motor current, door angle, opening and closing speed, and door operator output torque corresponding to each operating stage are collected. The motor current, door angle, opening and closing speed, and door operator output torque of each operating stage are used as control operation data. The ambient temperature, power supply voltage, door lock signal status, and control command response time of each operating stage are also collected. These parameters are used as auxiliary operation data. The time-domain statistical features of the control operation data and auxiliary operation data are extracted. The temporal statistical features of control operation data are used as control statistical features, and the temporal statistical features of auxiliary operation data are used as auxiliary statistical features to determine multiple types of features. By extracting control operation data and auxiliary operation data showing healthy door health from historical time points, control operation data showing healthy door health is identified as historical healthy operation data, and auxiliary operation data showing healthy door health is identified as historical healthy auxiliary data. These historical healthy operation data and historical healthy auxiliary data are used as health reference sample data. The temporal statistical features of the health reference sample data are extracted as health statistical features. At least one target cluster of health statistical features is determined using a target clustering algorithm. The mean and standard deviation of health statistical features for each dimension in each operational stage of the target cluster are used as cluster feature information. The product of the standard deviation and a preset parameter is determined as the first product. The difference between the mean and the first product corresponding to each cluster feature information is determined as the lower threshold. The sum of the mean and the first product corresponding to each cluster feature information is determined as the upper threshold. The lower and upper thresholds are combined to obtain the target detection threshold interval, realizing the dynamic determination of the target detection threshold interval. The target detection threshold interval of each control statistical feature is extracted as the first detection threshold, and the target detection threshold interval corresponding to each auxiliary statistical feature is extracted as the second detection threshold. The target detection threshold interval of each control statistical feature and its corresponding parameter are determined respectively. The system establishes a first attribution relationship for the first detection threshold and a second attribution relationship for each auxiliary statistical feature with the corresponding second detection threshold. Based on the first and second attribution relationships, the system determines the door health status. Control statistical features with a first attribution relationship of "not attributed" are designated as first control statistical features. Auxiliary statistical features with a second attribution relationship of "not attributed" are designated as second control statistical features. The combination of the first and second control statistical features is used as a feature combination. This feature combination is then matched against a preset fault database. Successfully matched fault patterns are designated as door fault patterns. This process accurately determines the door health status and door fault patterns, significantly improving the ability of the rail vehicle door system to identify structural faults.

[0094] Example 3

[0095] Figure 3This is a flowchart of a fault detection method according to Embodiment 3 of the present invention. This embodiment is a specific implementation of a fault detection method based on the above embodiments, such as... Figure 3 As shown, the method includes:

[0096] Step 1: Data collection and feature selection.

[0097] The collected data mainly includes raw signals (control operation data) reflecting the dynamic characteristics, motion state, and structural response of the door opening and closing process, including but not limited to motor current, door angle, opening and closing speed, and door operator output torque. Among these, motor current characterizes changes in the driving load, angle is used to locate the door's opening and closing amplitude and process, speed reflects the system's response speed, and torque is directly related to door obstruction and mechanism friction. Data collection is performed on different doors of the same vehicle model to ensure comparative analysis of subtle differences in opening and closing behavior under identical operating conditions. In addition, the system simultaneously collects a set of auxiliary operating parameters, including but not limited to ambient temperature, power supply voltage, door lock signal status, and control command response time, to describe the operating background and control environment. During the feature selection stage, this invention combines typical fault mode mechanisms of rail transit door systems (including but not limited to door misalignment, guide rail jamming, buffer structure wear, and abnormal motor braking) to systematically analyze and classify the collected raw signals, extracting the key features most strongly correlated with faults. In this embodiment, motor current and auxiliary operating parameters can be used as other features.

[0098] In one embodiment, to verify the applicability of the present invention in actual working conditions, a real-vehicle test was conducted on the door system of an in-service urban rail train. The test system is equipped with a standard door controller, an angle sensor, and a drive current acquisition module, and has the ability to acquire operating parameters such as motor current, door angle, opening and closing speed, and output torque in real time. At the same time, it acquires auxiliary information such as control signals, voltage, and response delay to form a complete state observation channel.

[0099] The system can use diesel engine operating data as input and focuses on typical structural faults in train doors. It constructs three representative test groups covering healthy, sub-healthy, and faulty states to ensure the comprehensiveness and comparability of the experiments. Each test group simulates the actual operation of the door system under different working conditions, including but not limited to: a healthy group (normal state without structural adjustments); a sub-healthy group (artificially simulating specific structural degradation problems, such as buffer head wear or guide rail jamming); and a faulty group (testing the extreme operating capacity of the door system through severe structural damage or fault simulation). Each test group performs no fewer than a specified number of opening and closing operations, with the opening and closing intervals controlled within a reasonable range to ensure coverage of typical peak and off-peak operating conditions during train operation. During sampling, the entire process of each opening and closing cycle is recorded and divided into a starting segment, a constant speed segment, a deceleration segment, and a final buffer segment for more accurate feature extraction and state analysis.

[0100] Step 2: Data feature extraction.

[0101] Stable operating phases with clearly defined fault-free markings are selected from historical operational data (including control and auxiliary operational data at historical moments) as health reference sample data. A health trend model reflecting typical relationships between key variables is constructed based on this sample set. Systematic feature extraction is performed on each type of collected characteristic signal to extract representative statistical features from the raw time-series data, comprehensively reflecting the dynamic behaviors and potential abnormal change trends during the door opening and closing process. Specifically, for each key parameter, multiple statistical indicators are extracted within a single opening and closing cycle, including but not limited to: mean, extreme values, rate of change, duration, and fluctuation amplitude. Based on the extracted statistical features, partitioned analysis can be performed by combining the segmented structure of the opening and closing process (such as the starting segment, constant speed segment, and deceleration segment) to further improve fault data localization. By constructing feature vectors for each feature dimension in each door action, a stable and usable state analysis feature dataset can be formed.

[0102] In one embodiment, the system can process data in batches. This invention, combining the physical mechanisms of typical structural faults in vehicle door systems, prioritizes the feature dimensions most sensitive to state changes, constructing a multi-dimensional feature set as input for subsequent modeling. Key features include, but are not limited to, parameters such as drive current, door angle, opening and closing speed, and output torque, which effectively reflect the stress state, mechanism response, and structural wear characteristics during the opening and closing process. In the feature extraction stage, the system performs segmented analysis on each opening and closing process from the original time-series data, extracting statistical indicators reflecting the evolution of the opening and closing state. These mainly include, but are not limited to, average current and its fluctuation amplitude, used to describe load changes; door angle change rate and stable plateau range, used to measure opening and closing uniformity; torque curve peak and continuously rising range, used to capture the mechanism's resistance performance; and closure stability and the degree of final buffer disturbance, used to identify buffer structure degradation characteristics. Simultaneously, by combining the time dimension, multiple stages of the opening and closing process are compared and analyzed, and a unified complete feature vector is constructed to achieve a comprehensive representation of a single opening and closing state, serving as the input basis for dynamic clustering and anomaly identification models.

[0103] Step 3: Based on K-means clustering modeling, extract statistical boundaries (i.e., upper and lower threshold limits) from each cluster to construct a threshold that is adaptively adjusted according to the door movement (i.e., the target detection threshold range).

[0104] K-means is a classic unsupervised clustering algorithm primarily used to divide a sample set into K clusters. Its core idea is to minimize the squared distance between each sample and its cluster center, thereby achieving optimal data grouping. The standard objective function of K-means is: ;in: Let be the feature vector of the i-th sample; It is the center of the k-th cluster; It is the sample set in the k-th cluster.

[0105] However, the standard K-means clustering objective only focuses on minimizing the distance between a sample and its cluster center to improve the separability between categories and the compactness of the clusters. However, this method does not constrain the degree of feature fluctuation within each cluster, which can easily lead to the following problems: (1) The distribution within the cluster is too loose: Although the samples in the same cluster are close to the center as a whole, their distribution range is wide, resulting in an excessively large dynamic threshold range, which reduces the sensitivity of anomaly identification; (2) The boundaries of the clustering results are blurred: When the fluctuation within the cluster is large, it is difficult to accurately define the boundary between "normal" and "abnormal", which affects the accuracy and stability of subsequent state discrimination; (3) It is not conducive to dynamic tolerance control: The lack of control over the distribution pattern of data within the cluster makes the dynamic threshold model based on clustering unstable in practical applications, which is difficult to meet the needs of refined diagnosis such as fault identification and trend monitoring.

[0106] Therefore, it is necessary to incorporate intra-cluster variance as a key indicator of intra-cluster structural stability into the clustering objective function for joint optimization. By limiting the "internal expansion" of each cluster, the fluctuation boundaries of features can be effectively tightened, thereby improving the discriminative power and robustness of the dynamic threshold model and achieving more reliable state recognition and anomaly detection. To this end, we introduce a new objective optimization function: The first term is to maintain the compactness of traditional K-means clustering, i.e., to minimize the sum of squared distances from a sample to its cluster center; the second term is to add the objective of "minimizing the variance of features within the cluster", which controls the dynamic threshold range. Used to balance the weights of the two parts. For value; The target number of initial clusters; This represents the k-th initial cluster; The number of samples in the k-th initial cluster; Health statistical features for a single health reference sample data (including multi-dimensional features such as motor current and door rotation angle). The center (mean) of the kth initial cluster; d is the squared Euclidean distance from the health statistics feature to the initial cluster center; d is the total number of dimensions of the health statistics feature (e.g., motor current, door rotation angle, opening and closing speed, and door machine output torque, a total of 4 features). Let the j-th feature be the value of the i-th health statistical feature; Let be the mean of the j-th feature of the k-th initial cluster; Let be the variance of the j-th feature of the k-th initial cluster.

[0107] The clustering structure and intra-cluster distribution parameters obtained based on the objective function optimization can provide constraint boundaries for the subsequent dynamic threshold model: in the k-th cluster, the fluctuation range of the j-th feature dimension is determined by its mean. and standard deviation Composition, for example: lower threshold Threshold upper limit .

[0108] In one embodiment, this invention employs a clustering and intra-class variance constraint mechanism to process a large amount of opening and closing process data collected during the actual operation of a subway car door system. This mechanism operates in an unsupervised manner, without relying on any manual labels or pre-classification information. First, the opening and closing feature data of multiple doors under the same car model are input into a clustering model for K-means clustering to divide the system into state clusters with similar characteristic behaviors. Unlike the traditional K-means method, this invention's model simultaneously introduces intra-class standard deviation constraints during the clustering process, performing statistical analysis and convergence control on the feature fluctuation range within each cluster, and constructing dynamic upper and lower bounds for each type of normal behavior.

[0109] The dynamic threshold model constructed in this invention can be continuously optimized as new data is introduced during the opening and closing of the car door, and has the ability to adapt to factors such as the door's action cycle, operating conditions, ambient temperature, and structural wear. Unlike traditional fixed threshold methods that rely on manual setting, this model automatically generates judgment boundaries based on the actual distribution patterns of features within clusters, and can dynamically adjust the threshold range, thereby maintaining higher recognition accuracy under varying operating conditions.

[0110] Step 4: Rule-based multi-feature collaborative fusion analysis mechanism.

[0111] After establishing a dynamic threshold (target detection threshold range), a multi-feature collaborative analysis mechanism based on rule and model fusion is constructed to parallelly fuse the discrimination results of multiple key features such as motor current, door rotation angle, opening and closing speed, and door operator output torque. By comprehensively considering the abnormal amplitude of each feature, the degree of deviation from the dynamic threshold, and the combination logic between them, a multi-dimensional assessment of the current operating state of the door is achieved, determining whether it is in a healthy, sub-healthy, or potentially faulty state. In one embodiment, a sub-healthy alarm can be triggered when the healthy state is sub-healthy, and a fault code can be triggered to push the fault mode when the healthy state is faulty.

[0112] Building upon this foundation, a fault triggering logic is further introduced: when multiple features simultaneously deviate from their normal fluctuation range and exhibit a specific combination pattern, the system automatically enters the fault mode identification stage. This stage matches the system with an established structured fault mode library, combining the temporal patterns and interrelationships of different feature offsets to accurately identify the typical problem type of the current fault. For example, if an abnormal increase in current is detected accompanied by a continuous increase in torque and a prolonged corner plateau period, it may indicate guide rail jamming; while a concurrent occurrence of corner offset and prolonged opening / closing time is more likely to indicate door offset or limit abnormality. Exemplary fault modes may include at least door offset, abnormal centering dimensions, guide rail wear, and buffer head wear.

[0113] In one embodiment, a multi-feature parallel analysis mechanism is used to achieve collaborative triggering and identification of structural fault modes. In this embodiment, the system, based on the constructed multi-feature parallel analysis mechanism, collaboratively judges key characteristic parameters during the door's operation, and combines the combination relationships between features to effectively distinguish different structural fault modes. Taking buffer head wear as an example, the system identifies the following characteristic combinations in multiple consecutive opening and closing cycles: a significant delay in the peak drive current, indicating a lag in load changes at the end of the closing phase; irregular high-frequency jitter in the torque signal at the buffer section, reflecting a decrease in energy absorption capacity at the end; and a slowdown in the door's corner closing rate, indicating unstable force feedback. This characteristic combination pattern highly matches the typical manifestations of fatigue wear in buffer structures. Based on this, the system triggers the buffer head wear identification logic, accurately distinguishing it from other fault types such as guide rail jamming and door misalignment, verifying the accuracy and mode adaptability of the proposed method in complex multi-feature scenarios.

[0114] This embodiment integrates multiple key operating parameters (such as motor current, speed, angle, and torque) for parallel discrimination to enhance the ability to detect structural faults. It combines cluster analysis to construct adaptive dynamic threshold boundaries, achieving high-precision classification of door states. This enables the identification and classification of specific fault modes, supporting more targeted maintenance decisions and state predictions, and providing technical support for intelligent train operation and maintenance. This invention significantly improves the ability of rail vehicle door systems to identify structural faults by integrating multi-feature parallel discrimination with clustering dynamic threshold modeling technology. Compared to traditional identification methods that rely on a single threshold and fixed logic, this invention can automatically adapt to different operating states and environmental changes, dynamically characterizing the normal fluctuation range of various key door parameters, enhancing the accuracy of identifying different fault modes such as buffer head wear and guide rail jamming, and helping to realize the transformation of rail vehicles from passive repair to proactive early warning.

[0115] Example 4

[0116] Figure 4 This is a schematic diagram of a fault detection device according to Embodiment 4 of the present invention. Figure 4 As shown, the device includes: a feature acquisition module 41, a feature determination module 42, a range determination module 43, and a fault detection module 44.

[0117] The feature acquisition module 41 is used to acquire the control operation data and auxiliary operation data of the door system, and to determine the control statistical features of the control operation data and the auxiliary statistical features of the auxiliary operation data.

[0118] The feature determination module 42 is used to acquire historical health operation data and historical health auxiliary data as health reference sample data, and to determine the health statistical features of the health reference sample data.

[0119] The interval determination module 43 is used to determine the target cluster of health statistical features through the target clustering algorithm, and to determine the target detection threshold interval of control statistical features and auxiliary statistical features based on the cluster feature information of the target cluster.

[0120] The fault detection module 44 is used to determine the health status and fault mode of the door based on control statistical features, auxiliary statistical features and target detection threshold range.

[0121] The technical solution of this invention involves a feature acquisition module that obtains control operation data and auxiliary operation data of the door system, determines the control statistical features of the control operation data and the auxiliary statistical features of the auxiliary operation data, a feature determination module that obtains historical health operation data and historical health auxiliary data as health reference sample data, determines the health statistical features of the health reference sample data, and determines the target cluster of the health statistical features through a target clustering algorithm, an interval determination module that determines the target detection threshold interval of the control statistical features and auxiliary statistical features based on the cluster feature information of the target cluster, and realizes the dynamic construction of the target detection threshold interval, and a fault detection module that determines the door health status and door fault mode based on the control statistical features, auxiliary statistical features and target detection threshold interval, and integrates multiple control operation data and auxiliary operation data for parallel discrimination to improve the perception capability of structural faults, realize the accurate determination of the door health status and door fault mode, and improve the efficiency of determining the door health status and door fault mode.

[0122] In one embodiment, the fault detection device further includes:

[0123] The candidate data determination module is used to determine control operation data and auxiliary operation data that indicate a healthy door health status as candidate healthy operation data, and to determine auxiliary operation data that indicates a healthy door health status as candidate healthy auxiliary data.

[0124] The candidate sample determination module is used to use candidate health operation data and candidate health auxiliary data as candidate health reference sample data.

[0125] The data update module is used to use candidate health reference sample data as health reference sample data when the number of candidate health reference sample data exceeds a preset threshold.

[0126] In one embodiment, the feature acquisition module 41 includes:

[0127] The stage determination unit is used to collect the rotational speed of the door system during the opening and closing process, and divide the opening and closing process into at least one operating stage according to the rotational speed.

[0128] The first data determination unit is used to collect the motor current, door rotation angle, opening and closing speed and door machine output torque corresponding to each operating stage, and use the motor current, door rotation angle, opening and closing speed and door machine output torque of each operating stage as control operation data.

[0129] The second data determination unit is used to collect the ambient temperature, power supply voltage, door lock signal status and control command response time of each operating stage, and use the ambient temperature, power supply voltage, door lock signal status and control command response time of each operating stage as auxiliary operating data.

[0130] The feature extraction unit is used to extract the time-domain statistical features of control operation data and auxiliary operation data respectively, using the time-domain statistical features of control operation data as control statistical features and the time-domain statistical features of auxiliary operation data as auxiliary statistical features.

[0131] In one embodiment, the feature determination module 42 includes:

[0132] The historical data extraction unit is used to extract control operation data and auxiliary operation data of the door health status in historical moments.

[0133] The sample determination unit is used to determine the control operation data of the door health status as healthy as historical health operation data, the auxiliary operation data of the door health status as healthy as historical health auxiliary data, and the historical health operation data and historical health auxiliary data as health reference sample data.

[0134] The health feature determination unit is used to extract the time-domain statistical features of the health reference sample data as health statistical features.

[0135] In one embodiment, the interval determination module 43 includes:

[0136] Clustering cluster determination unit, used to determine at least one target cluster of health statistical characteristics through a target clustering algorithm;

[0137] The cluster feature determination unit is used to determine the mean and standard deviation of the health statistical features of each dimension in each running stage of the target cluster, and uses the mean and standard deviation as cluster feature information.

[0138] The threshold interval determination unit is used to determine the product of the standard deviation and the preset parameter as the first product, the difference between the mean of each cluster feature information and the first product as the lower threshold, the sum of the mean of each cluster feature information and the first product as the upper threshold, and the combination of the lower threshold and the upper threshold to obtain the target detection threshold interval.

[0139] In one embodiment, the cluster determination unit is specifically used for:

[0140] Determine the cluster centers of the initial clusters for the target number of health statistical features, determine the squared distance between each health statistical feature and each cluster center as the distance value, and determine the sum of the variances of each dimension of the health statistical features in the initial clusters in which the health statistical features are located as the variance value.

[0141] The distance and variance values ​​are input into the objective optimization function to determine the cost value, and the initial cluster with the minimum cost value is used as the intermediate cluster for each health statistical feature.

[0142] Determine the mean of the health statistical features in each intermediate cluster, use the mean of the health statistical features as the target cluster center, redetermine the intermediate clusters of each health statistical feature according to the target cluster center, until the rate of change of the cost value is less than the target value, and determine the intermediate clusters corresponding to each health statistical feature as the target clusters.

[0143] In one embodiment, the fault detection module 44 includes:

[0144] The threshold allocation unit is used to extract the target detection threshold interval of each control statistical feature as the first detection threshold, and extract the target detection threshold interval corresponding to each auxiliary statistical feature as the second detection threshold.

[0145] The health status determination unit is used to determine the first attribution relationship between each control statistical feature and the corresponding first detection threshold and the second attribution relationship between each auxiliary statistical feature and the corresponding second detection threshold, and to determine the health status of the door according to the first attribution relationship and the second attribution relationship.

[0146] The combination determination unit is used to determine the control statistical feature that is not assigned to the first attribution relationship as the first control statistical feature, determine the auxiliary statistical feature that is not assigned to the second attribution relationship as the second control statistical feature, and take the combination of the first control statistical feature and the second control statistical feature as the feature combination.

[0147] The fault mode determination unit is used to match the fault mode in the preset fault database according to the feature combination, and the successfully matched fault mode is taken as the door fault mode.

[0148] The fault detection device provided in the embodiments of the present invention can execute the fault detection method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the method.

[0149] Example 5

[0150] Figure 5 This is a schematic diagram of the structure of an electronic device implementing the fault detection method of an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0151] like Figure 5 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0152] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0153] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as a fault detection method.

[0154] In some embodiments, a fault detection method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of a fault detection method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform a fault detection method by any other suitable means (e.g., by means of firmware).

[0155] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0156] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0157] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0158] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0159] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0160] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0161] In one embodiment, the present invention further includes a computer program product, which includes a computer program that, when executed by a processor, implements a fault detection method according to any embodiment of the present invention.

[0162] In implementing the computer program product, computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0163] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0164] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A fault detection method, characterized in that, include: The door system's rotational speed during the opening and closing process is collected, and the opening and closing process is divided into at least one operating stage according to the rotational speed. The motor current, door angle, opening and closing speed and door machine output torque corresponding to each of the above-mentioned operating stages are collected respectively, and the motor current, door angle, opening and closing speed and door machine output torque of each operating stage are used as control operation data; Ambient temperature, power supply voltage, door lock signal status, and control command response time are collected for each of the aforementioned operating stages, and the ambient temperature, power supply voltage, door lock signal status, and control command response time for each operating stage are used as auxiliary operating data. The time-domain statistical features of the control operation data and the auxiliary operation data are extracted respectively. The time-domain statistical features of the control operation data are used as control statistical features, and the time-domain statistical features of the auxiliary operation data are used as auxiliary statistical features. Historical health operation data and historical health auxiliary data are obtained as health reference sample data, and the health statistical characteristics of the health reference sample data are determined. The target clusters of the health statistical features are determined by the target clustering algorithm, and the target detection threshold ranges of the control statistical features and the auxiliary statistical features are determined based on the cluster feature information of the target clusters. The door health status and door failure mode are determined based on the control statistical features, the auxiliary statistical features, and the target detection threshold range. The step of determining the target clusters of the health statistical features using a target clustering algorithm, and determining the target detection threshold ranges of the control statistical features and the auxiliary statistical features based on the cluster feature information of the target clusters, includes: Determine the cluster centers of the initial clusters for the target number of health statistical features, determine the squared distance between each health statistical feature and each cluster center as the distance value, and determine the sum of the variances of each dimension of the health statistical features in the initial clusters in which the health statistical features are located as the variance value. The distance value and variance value are input into the objective optimization function to determine the cost value, and the initial cluster with the smallest cost value is used as the intermediate cluster for each health statistical feature. The mean of the health statistical features in each intermediate cluster is determined, and the mean of the health statistical features is used as the target cluster center. The intermediate clusters of each health statistical feature are re-determined according to the target cluster center until the rate of change of the cost value is less than the target value. The intermediate clusters corresponding to each health statistical feature are then determined as the target clusters. Determine the mean and standard deviation of the health statistical features of each dimension in each operational stage of the target cluster, and use the mean and standard deviation as cluster feature information; The product of the standard deviation and the preset parameter is determined as the first product. The difference between the mean of each cluster feature information and the first product is determined as the lower threshold. The sum of the mean of each cluster feature information and the first product is determined as the upper threshold. The lower threshold and the upper threshold are combined to obtain the target detection threshold range. The step of determining the door health status and door fault mode based on the control statistical features, the auxiliary statistical features, and the target detection threshold range includes: The target detection threshold intervals of each control statistical feature are extracted as the first detection threshold, and the target detection threshold intervals corresponding to each auxiliary statistical feature are extracted as the second detection threshold. The first attribution relationship between each control statistical feature and its corresponding first detection threshold and the second attribution relationship between each auxiliary statistical feature and its corresponding second detection threshold are determined respectively. The health status of the car door is determined according to the first attribution relationship and the second attribution relationship. The control statistical feature that determines the first attribution relationship as non-attribution is used as the first control statistical feature, and the auxiliary statistical feature that determines the second attribution relationship as non-attribution is used as the second control statistical feature. The combination of the first control statistical feature and the second control statistical feature is used as the feature combination. The fault patterns are matched against a preset fault database based on the combination of features, and the successfully matched fault patterns are taken as the door fault patterns.

2. The method according to claim 1, characterized in that, After determining the door health status and fault mode based on the control statistical features, the auxiliary statistical features, and the target detection threshold range, the method further includes: Control operation data that determines the health status of the vehicle door to be healthy is used as candidate health operation data, and auxiliary operation data that determines the health status of the vehicle door to be healthy is used as candidate health auxiliary data; The candidate health operation data and the candidate health auxiliary data are used as candidate health reference sample data; When the number of candidate health reference sample data exceeds a preset threshold, the candidate health reference sample data is used as health reference sample data.

3. The method according to claim 1, characterized in that, The process of acquiring historical health operation data and historical health auxiliary data as health reference sample data, and determining the health statistical characteristics of the health reference sample data, includes: Extract control operation data and auxiliary operation data of healthy door status from historical time points; The control operation data that determines the health status of the vehicle door to be healthy is used as historical health operation data, and the auxiliary operation data that determines the health status of the vehicle door to be healthy is used as historical health auxiliary data. The historical health operation data and the historical health auxiliary data are used as health reference sample data. The time-domain statistical features of the health reference sample data are extracted as health statistical features.

4. A fault detection device, characterized in that, include: The stage determination unit is used to collect the rotational speed of the door system during the opening and closing process, and divide the opening and closing process into at least one operating stage according to the rotational speed. The first data determination unit is used to collect the motor current, door rotation angle, opening and closing speed and door machine output torque corresponding to each of the operation stages, and use the motor current, door rotation angle, opening and closing speed and door machine output torque of each operation stage as control operation data. The second data determination unit is used to collect the ambient temperature, power supply voltage, door lock signal status and control command response time of each of the aforementioned operating stages, and use the ambient temperature, power supply voltage, door lock signal status and control command response time of each operating stage as auxiliary operating data. The feature extraction unit is used to extract the time-domain statistical features of the control operation data and the auxiliary operation data respectively, and use the time-domain statistical features of the control operation data as control statistical features and the time-domain statistical features of the auxiliary operation data as auxiliary statistical features. The feature determination module is used to acquire historical health operation data and historical health auxiliary data as health reference sample data, and to determine the health statistical features of the health reference sample data. The interval determination module is used to determine the target cluster of the health statistical features through a target clustering algorithm, and to determine the target detection threshold interval of the control statistical features and the auxiliary statistical features based on the cluster feature information of the target cluster. The fault detection module is used to determine the door health status and door fault mode based on the control statistical features, the auxiliary statistical features and the target detection threshold range; The interval determination module is used for: Determine the cluster centers of the initial clusters for the target number of health statistical features, determine the squared distance between each health statistical feature and each cluster center as the distance value, and determine the sum of the variances of each dimension of the health statistical features in the initial clusters in which the health statistical features are located as the variance value. The distance value and variance value are input into the objective optimization function to determine the cost value, and the initial cluster with the smallest cost value is used as the intermediate cluster for each health statistical feature. The mean of the health statistical features in each intermediate cluster is determined, and the mean of the health statistical features is used as the target cluster center. The intermediate clusters of each health statistical feature are re-determined according to the target cluster center until the rate of change of the cost value is less than the target value. The intermediate clusters corresponding to each health statistical feature are then determined as the target clusters. Determine the mean and standard deviation of the health statistical features of each dimension in each operational stage of the target cluster, and use the mean and standard deviation as cluster feature information; The product of the standard deviation and the preset parameter is determined as the first product. The difference between the mean of each cluster feature information and the first product is determined as the lower threshold. The sum of the mean of each cluster feature information and the first product is determined as the upper threshold. The lower threshold and the upper threshold are combined to obtain the target detection threshold range. The fault detection module includes: The threshold allocation unit is used to extract the target detection threshold interval of each control statistical feature as the first detection threshold, and extract the target detection threshold interval corresponding to each auxiliary statistical feature as the second detection threshold. The health status determination unit is used to determine the first attribution relationship between each control statistical feature and the corresponding first detection threshold and the second attribution relationship between each auxiliary statistical feature and the corresponding second detection threshold, and to determine the health status of the door according to the first attribution relationship and the second attribution relationship; The combination determination unit is used to determine the control statistical feature that is not assigned to the first attribution relationship as the first control statistical feature, determine the auxiliary statistical feature that is not assigned to the second attribution relationship as the second control statistical feature, and take the combination of the first control statistical feature and the second control statistical feature as the feature combination. The fault mode determination unit is used to match the fault mode in a preset fault database according to the combination of features, and to take the successfully matched fault mode as the door fault mode.

5. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the fault detection method according to any one of claims 1-4.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the fault detection method according to any one of claims 1-4.

Images