A method and system for evaluating the installation and adjustment quality of rail transit vehicle doors and a rail vehicle

Abstract

The present invention discloses a method and system for evaluating the assembly and adjustment quality of rail transit doors and a rail vehicle. By using a rail transit door assembly and adjustment quality evaluation system with edge computing capabilities and combining machine learning algorithms, the running state curves of the doors collected are clustered, analyzed, and matched to screen out the doors in unqualified assembly states and output the types and causes of unqualified states; for the doors in qualified assembly states, the assembly quality of the doors is quantitatively displayed. This solution improves the intelligent level and installation efficiency of door installation; data processing is realized locally, eliminating the cumbersome process of data transmission and the process of diagnosing data on the background server, and greatly improving the efficiency of data processing.

Description

Technical Field

[0001] The present invention belongs to the field of urban rail transit, and particularly relates to a method and system for evaluating the installation and adjustment quality of rail transit doors and a rail vehicle. Background Art

[0002] With the development of the rail transit industry, the safety and reliability of the operation of urban rail transit doors have received extensive attention. The complex structure of the door system makes it easy to have various defects of unqualified installation during the installation process of the doors. Therefore, a high qualification rate when the doors are installed, adjusted and taken off the production line is of great significance for improving the operation quality and reliability of the door system, and even for the efficient and safe operation of urban rail transit trains.

[0003] The existing door debugging devices can only simply read and judge whether the fault codes, door opening and closing signals and the I / O signals of the door controller are normal, and cannot guide the installation and adjustment of the doors, nor can they evaluate the quality of the door installation. Summary of the Invention

[0004] Object of the Invention: The first object of the present invention is to provide a method for evaluating the installation and adjustment quality of rail transit doors; the second object of the present invention is to provide a system for evaluating the installation and adjustment quality of rail transit doors; the third object of the present invention is to provide a rail vehicle with a function of evaluating the installation and adjustment quality of doors.

[0005] Technical Solution: A method for evaluating the installation and adjustment quality of rail transit doors according to the present invention includes the following steps:

[0006] (1) Obtain the operation curve data of not less than one door after installation and adjustment;

[0007] (2) Compare and analyze the operation curve data of each door after installation and adjustment with the standard operation curve data of the door to determine whether the assembly quality of the door is qualified; when it is determined that the assembly quality of the door is unqualified, output the relevant data of the unqualified door assembly.

[0008] Preferably, the relevant data of the unqualified door assembly output in step (2) are the unqualified status category of the door and / or the name of the unqualified door assembly component.

[0009] Preferably, in step (2), the doors with qualified assembly quality are scored for the assembly quality, and the doors that meet the requirements according to the set rules are screened.

[0010] Preferably, in step (2), the operation curve features of the doors after installation and adjustment are extracted, combined with a preset operation curve feature database, and the trained machine learning algorithm is used to complete the determination of the door assembly quality. The operation curve feature database includes the operation curve feature data of the doors with qualified assembly and the operation curve feature data of the doors in an unqualified assembly state.

[0011] Preferably, after extracting and adjusting the running curve characteristics of the door in step (2), the determination of whether the door assembly quality is qualified is completed through cluster analysis in combination with the running curve characteristic data of the qualified assembled door; when the door assembly quality after adjustment is determined to be unqualified, the running curve characteristics of the door are matched with the running curve characteristic data of the doors in the unqualified assembly state in the curve characteristic database to complete the determination of the unqualified state category of the door.

[0012] Preferably, when the door assembly is in a qualified state, the assembly quality score of the door is calculated based on the deviation amount of the door running curve characteristic value relative to the standard running curve characteristic value.

[0013] Furthermore, in step (1), the door running curve data within a set time range is obtained, such as the running curve data of 50 or 100 times of the door opening and closing process. The door running curve data includes the motor current curve, speed curve, and rotation angle curve.

[0014] Furthermore, when constructing the preset running curve characteristic database in step (2), first, the running state curves of the doors in different unqualified installation states or the doors with unqualified installation parts are collected. After preprocessing the collected data, it is compared and analyzed with the standard running state curve data of the standard door type. Through steps such as statistical eigenvalue, data standardization processing, data discreteness weight analysis, and cluster analysis, the unique change characteristics of the running state curves of each door with unqualified assembly quality are obtained.

[0015] Furthermore, after constructing the running curve characteristic database, the door running curve data collected on-site can be compared and analyzed. In this process, the running curve data of a single door can be collected, or the running curve data of several doors can be monitored and collected simultaneously for individual analysis.

[0016] When comparing and analyzing the door running curve data collected on-site, first, cluster analysis is performed in combination with the standard running state curve data. If the clustering fails, it is determined that the door is in a qualified state; if the clustering is successful, it indicates that the installation and adjustment of the currently detected door are in an unqualified state. After extracting the eigenvalue from the running state curve data of the door, it is compared with the data in the running curve characteristic database. If the trend characteristic is within the set threshold range of any data in the database, the unqualified state of the installation and adjustment of the currently detected door can be determined.

[0017] Furthermore, for each unqualified assembly state in the running curve characteristic database, there is a corresponding solution. After the determination of the unqualified state of the currently detected door is completed, the solution for this unqualified state is matched and output to help the installation and adjustment personnel perform the debugging and installation of the door.

[0018] Further, for a qualified assembled door, the eigenvalue of its operation curve is within a set threshold range relative to the standard eigenvalue. This solution can display the doors that have been debugged and installed qualified through quantitative analysis. The maximum deviation amplitude of the relative standard eigenvalue among all eigenvalues is used as the basis for quantitative analysis. For example: If all the calculated eigenvalues of the doors are consistent with the eigenvalues of the standard door type, the installation quality of the door is rated 100 points, indicating that the door is perfectly installed without any deviation in any component; a threshold is set based on each eigenvalue of the standard door type. If the eigenvalue with the maximum deviation amplitude is measured to fluctuate within 1% of its standard eigenvalue, it is determined that the installation quality of the door is between 100 points and 90 points; if the eigenvalue with the maximum deviation amplitude is measured to exceed 1% but not exceed 2% of its standard eigenvalue, it is considered that the installation quality of the door is between 90 points and 80 points; and so on. If the eigenvalue with the maximum deviation amplitude is measured to exceed 5% of its standard eigenvalue, it is directly determined that the door installation is unqualified and needs to be reinstalled and adjusted. The specific score can be obtained through ratio operations based on the specific deviation amplitude.

[0019] This method of quantitatively displaying the installation and adjustment quality of the doors that have been debugged and installed qualified can rank the assembly quality of each qualified door, which is beneficial to marking the doors that, although in a qualified assembly state, deviate more from the standard state, providing a basis for subsequent key monitoring of the doors.

[0020] A rail transit door installation and adjustment quality evaluation system according to the present invention includes a door parameter analysis device with edge computing function and a data acquisition unit. The data acquisition unit is connected to the door parameter analysis device, and the door parameter analysis device is used to execute the rail transit door installation and adjustment quality evaluation method described in any one of the above.

[0021] Preferably, the door parameter analysis device is used to output the evaluation result of the door assembly quality. The door parameter analysis device is connected to a human-machine interaction screen, and the human-machine interaction screen is used to display the evaluation result of the door assembly quality. The door parameter analysis device is connected to an external terminal through a communication unit.

[0022] A rail vehicle according to the present invention includes a door controller and a door parameter analysis device with edge computing function. The door controller is connected to the door parameter analysis device, and the door parameter analysis device is used to execute the rail transit door installation and adjustment quality evaluation method described in any one of the above.

[0023] Preferably, the door controller includes a main door controller and several slave door controllers for collecting the operation state data of individual doors. The main door controller and the slave door controllers form a communication ring network, and the main door controller is connected to the door parameter analysis device.

[0024] Furthermore, the door parameter analysis device with edge computing function adopts the high-performance processor RK3568, 16GB large-capacity DDR4 and 64GB eMMC to form the hardware processing unit required by the system, which can process intelligent learning algorithms such as machine learning. It uses RS485, RS232, CAN, ETH, MVB and other wired communication interfaces as data acquisition units. Under the premise of not changing the original door controller network, the door parameter analysis device can be directly connected to the door controller network of each car to obtain the door opening and closing data of each car.

[0025] In addition, the door parameter analysis device is also connected to a variety of wireless network interfaces, such as 3G, 4G, 5G, WIFI, ZigBee, etc.; the door parameter analysis results are sent to other terminals or background servers, and the system can be remotely upgraded and the algorithm updated using the wireless network interface. In the door assembly quality evaluation results displayed on the human-computer interaction screen, for doors that pass the assembly test, the assembly quality score is output, and for doors that fail the assembly test, the type of unqualified state and the name of the unqualified assembly parts are output.

[0026] Furthermore, the door parameter analysis device has edge computing capabilities and can be installed locally on the door or made into a miniaturized device that is easy to carry. Therefore, the data can be analyzed and processed directly locally to guide the installation of the door. There is no need to transmit the data to a ground server for diagnosis, and it is not constrained by the in-vehicle data transmission system. There is no need to change the original data transmission method, and there is almost no change to the door transmission system.

[0027] Furthermore, in a rail vehicle, a door controller group in the same car or the same train of rail vehicles adopts ring network communication, and the door parameter analysis device only needs to be connected to one door controller to collect the operating parameters of all door controllers, which can greatly improve the detection efficiency.

[0028] Beneficial effects: The technical solution disclosed in the present invention mines, analyzes and processes the door operation status curve data through an artificial intelligence diagnostic algorithm to achieve installation guidance and installation quality evaluation for the door, which greatly improves the intelligence level of door installation and also greatly improves the efficiency of door installation.

[0029] Furthermore, this detection method and system can realize localized data processing. The system is portable and utilizes edge computing and cloud-edge collaboration technologies to eliminate the tedious process of data transmission and the process of data diagnosis on the background server, greatly improving the efficiency of data processing. It can not only improve the efficiency of vehicle door installation quality evaluation, but also improve the stability of the evaluation process system. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1It is a schematic diagram of the rail vehicle and door assembly and adjustment quality evaluation system in the present invention;

[0031] Figure 2 It is a schematic diagram of the structure of the door assembly and adjustment quality evaluation system in the present invention;

[0032] Figure 3 It is the touch LCD screen door installation guidance interface in the present invention;

[0033] Figure 4 It is a flow chart of the rail transit door assembly and adjustment quality evaluation method in the present invention;

[0034] Figure 5 It is a flow chart of constructing the operating curve feature database in the present invention;

[0035] Figure 6 It is a flow chart of comparing and analyzing the door operating curve data collected on site in the present invention. Specific implementation mode

[0036] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0037] Taking the door assembly and adjustment quality evaluation of the plug and slide door system commonly used in urban rail transit as an example, as Figure 1 shown, the rail vehicle 101 includes a train 102, a main door controller 103, slave door controllers 104 - 110, a rail transit door assembly and adjustment quality evaluation system 111, a background server 112, and other portable device terminals 113. The rail transit door assembly and adjustment quality evaluation system is connected to the main door controller through a wired communication interface. The main door controller is connected to each slave door controller to form a ring network. Each slave door controller can send the door opening and closing data to the main door controller uniformly. After receiving the door opening and closing data of the corresponding door sent by each door controller, the rail transit door assembly and adjustment quality evaluation system conducts door assembly and adjustment quality evaluation, and can display the result locally, or can send the prediction result or door data to the background server through wireless networks such as 3G / 4G / 5G for storage and further analysis and processing, and can also send it to other portable device terminals through short-range wireless such as WIFI, BlueTooth, ZigBee for display and storage.

[0038] In this embodiment, the rail transit door assembly and adjustment quality evaluation system includes a door parameter analysis device with edge computing function, as Figure 2As shown, the door parameter analysis device consists of a high-performance processor RK3568, 16GB of large-capacity DDR4, and 64GB of eMMC. The rail transit door installation and adjustment quality evaluation system also includes a wired communication interface, a power supply conversion system, a rechargeable battery, a Type-C charging circuit, a power conversion circuit, a 3G / 4G / 5G module, a ZigBee module, a WIFI&BlueTooth module, a USB3.0 interface circuit, and a touch LCD screen. Among them, the wired communication interface circuit consists of an RS485 communication interface, an RS232 communication interface, a CAN communication interface, an ETH communication interface, and an MVB communication interface circuit. The power conversion circuit is composed of power chip conversion 1, power chip conversion 2, and power chip conversion 3. The wired communication interface is used to adapt to any interface of the main door controller, facilitating connection to the main door controller network. The system is powered by in-vehicle 110VDC, which can be converted into the voltage required by the system after passing through the power supply conversion system. At the same time, this portable device can also be powered by a rechargeable battery, and the rechargeable battery can be powered by the Type-C charging light. After the power supply system converts the 110VDC voltage into the 5.2V voltage required by the system, it is then converted into a 1.21V voltage by power chip conversion 1 to supply power to DDR4. Power conversion chip 1 simultaneously converts 1.8V, 3.3V, and 1.0V voltages to supply power to the main control chip. Power conversion chip 2 generates 3.3V and 5.0V voltages to supply power to eMMC, the ZigBee module, and the WIFI&BlueTooth module. Power conversion chip 3 generates 3.8V voltage to supply power to the 3G / 4G / 5G module. The touch LCD screen is connected to the main control chip through an RGB interface, and the touch interface is IIC; the 3G / 4G / 5G module is connected to the main controller through a USB3.0 interface; the ZigBee module communicates with the main controller through a UART interface; the WIFI&BlueTooth module communicates with the main controller through an SDMMC1 interface.

[0039] In this embodiment, after the wired communication interface is connected to the main door controller network, the rail transit door assembly and adjustment quality evaluation system sends a handshake command 0x01 0x02 0x03 0xff to the main door controller. At this time, the main door controller replies 0xff 0x03 0x02 0x01. Then the rail transit door assembly and adjustment quality evaluation system sends another handshake command 0x11 0x12 0x13 0xff to the main door controller. At this time, the main door controller replies 0xff 0x13 0x12 0x11. After the two handshakes are successful, it indicates that the rail transit door assembly and adjustment quality evaluation system has been successfully connected to the main door controller. Since the main door controller is connected to each slave door controller to form a ring network, each slave door controller sends the door opening and closing data to the main door controller uniformly, and then the main door controller sends all the door opening and closing data collected in this carriage to a kind of rail transit door assembly and adjustment quality evaluation system device. The evaluation algorithms such as the K-means algorithm and the artificial intelligence diagnosis algorithm are pre-stored in the eMMC of the rail transit door assembly and adjustment quality evaluation system. For the collected door opening and closing data, relevant algorithms are called to output the evaluation result of the door assembly and adjustment quality. Finally, according to the processing result, the unqualified doors are further assembled and adjusted.

[0040] In this embodiment, the evaluation result of the door assembly and adjustment quality can be directly displayed on the local LCD screen, such as Figure 3 shown. After the evaluation of the door assembly and adjustment quality, the installation quality evaluation scores of each door and the door numbers that need to be readjusted will appear on the door installation guidance prediction page. Taking the two unqualified assembled doors 113A and 114B as an example, when a specific door number is selected, such as 113A, it will jump to the door interface that needs to be readjusted, and at this time, the components that need to be readjusted for this door will be prompted. If 114B is selected, it will jump to the door interface that needs to be readjusted, and at this time, the components that need to be readjusted for this door will be prompted.

[0041] In this embodiment, when the artificial intelligence diagnosis algorithm pre-stored in the eMMC of the rail transit door assembly and adjustment quality evaluation system is used to evaluate the rail transit door assembly and adjustment quality, the specific method steps are as Figure 4 shown. The specific steps are as follows:

[0042] (1) Send door opening and closing commands to each door through the connected wired communication interface, so as to control the doors to perform 50 or a specified number of door opening and closing experiments. Each door packs and transmits back the operation curve data such as the motor current curve, speed curve, and rotation angle curve during the door opening and closing process and the door number uniformly.

[0043] (2) The system compares the collected door operation curve with the built-in standard curve of the door, uses the K-means algorithm to extract the eigenvalues of the two curves and analyzes according to the eigenvalues. When the clustering analysis result of the eigenvalues of the collected door operation curve and the standard curve eigenvalues fails, it is determined that the door is in a qualified assembly state. When the clustering analysis result is successful, it is determined that the door is in an unqualified assembly state;

[0044] (3) For the door in the unqualified assembly state, after being processed by the K-means algorithm, the extracted eigenvalues are matched with the data in the preset operation curve eigenvalue database. When it matches any unqualified type, such as abnormal screw rod lubrication, abnormal nut, worn buffer head, abnormal lower stop pin, etc., the system will alarm on the LCD screen, indicating the unqualified installation state of a specific door and the unqualified parts of the door installation;

[0045] (4) For the door detected to be in the qualified state, the debugging and installation quality is quantitatively displayed, and the maximum deviation amplitude of the relative standard eigenvalue among all eigenvalues is used as the basis for quantitative analysis. In this embodiment, if all the door eigenvalues actually calculated are consistent with the eigenvalues of the standard door type, the installation quality of the door is set to 100 points, indicating that the door is installed perfectly and there is no deviation in any component; A threshold is set based on each eigenvalue of the standard door type. If the eigenvalue with the measured maximum deviation amplitude fluctuates within 1% of its standard eigenvalue, it is determined that the installation quality of the door is between 100 points and 90 points; If the eigenvalue with the measured maximum deviation amplitude exceeds 1% of its standard eigenvalue and does not exceed 2%, it is considered that the installation quality of the door is between 90 points and 80 points; And so on. If the eigenvalue with the measured maximum deviation amplitude exceeds 5% of its standard eigenvalue, it is directly determined that the door installation is unqualified and needs to be reinstalled. The specific score can be obtained through specific ratio operations of the deviation amplitude.

[0046] In this embodiment, when constructing the operation curve eigenvalue database, the operation curves of the doors in various unqualified installation states are modeled. First, the normal operation curve of the standard door type is collected as the benchmark. During the modeling process, for the known unqualified installation states, the door operation curve data at this time is collected and compared and analyzed with the benchmark data. As Figure 5 shown, all the change characteristics of this unqualified assembly state are found, that is, the operation curve eigenvalue model of the unqualified assembly door. Model all the known unqualified states of door assembly to generate the operation curve eigenvalue database.

[0047] In this embodiment, the K-means algorithm is used to extract the characteristic values of the operation state curve. The opening and closing operation curve of the car door is divided into five stages: acceleration stage, high-speed stage, deceleration stage, slow-moving stage, and post-arrival stage. The extraction of characteristic values mainly includes the following seven parameters: mean, range, variance, standard deviation, skewness, kurtosis, and correlation coefficient. Calculating the characteristic values means calculating the mean, range, variance, standard deviation, skewness, kurtosis, and correlation coefficient of each stage in the five stages of the opening and closing operation curve of the car door. Taking the poor installation of the lower stop pin as an example to illustrate the execution process of the K-means algorithm: First, collect the operation curve of the standard door type, and use the K-means algorithm to find the seven characteristic values of each stage in the process of opening and closing the door; then collect the operation curve of the door type with poor installation of the lower stop pin, and find the seven characteristic values of each stage in the process of opening and closing the door. Cluster and analyze the characteristic values of the two groups of data to obtain a rule for poor installation of the lower stop pin component, and this rule is recorded as the rule for poor installation of the lower stop pin. When conducting car door installation guidance and evaluation, it is to collect the actual car door operation curve data on site, match the characteristic values extracted from this data with the characteristic values of poor installation of the lower stop pin component. If the trend characteristics are within a certain threshold, it can be inferred that there is a situation of poor installation of the lower stop pin component installed on this door type.

[0048] In this embodiment, for the car doors in a qualified state of assembly, the assembly quality is scored by the degree of deviation of the maximum characteristic value. The car doors with an assembly quality score close to 60 points or several car doors with the lowest assembly quality scores are marked to provide the car door data that needs to be focused on during the subsequent maintenance of the vehicle operation.

[0049] To sum up, the rail transit car door assembly and adjustment quality evaluation system with edge computing function combines machine learning algorithms to call the pre-built standard door type operation curve and compare it with the obtained actual car door curve. The characteristic values at the difference between the standard operation curve and the actual operation curve are extracted through the K-meas algorithm to judge the unqualified state and defective components of the car door installation, and to guide the installation and debugging work of the car door, thus realizing the evaluation of the car door installation quality, which greatly improves the efficiency of car door installation. The rail transit car door assembly and adjustment quality evaluation system with edge computing function directly analyzes and processes the data locally to guide the car door installation, without the need to transmit the data to the ground server for diagnosis, is not restricted by the in-vehicle data transmission system, and does not need to change the original data transmission method, simplifies the system complexity, and improves the system operation efficiency and operation stability. By diagnosing the assembly faults of unqualified assembled car doors and quantitatively displaying the assembly quality of qualified assembled car doors, a complete car door assembly quality evaluation system is formed, which can effectively improve the passing rate of car door assembly production offline and also provide support data for subsequent operation and maintenance.

Claims

1. A method for evaluating the installation and adjustment quality of rail transit doors, characterized in that: The method includes the following steps: (1) Obtain the operation curve data of at least one door after installation and adjustment; (2) Compare and analyze the operation curve data of each door after installation and adjustment with the standard operation curve data of the door to determine whether the assembly quality of the door is qualified; When it is determined that the assembly quality of the door is unqualified, output the relevant data of the unqualified door assembly; When it is determined that the assembly quality of the door is qualified, score the assembly quality of the door whose assembly quality is determined to be qualified, screen the doors that meet the requirements according to the set rules; extract the operation curve characteristics of the door after installation and adjustment, combine the preset operation curve characteristic database, and use the trained machine learning algorithm to complete the determination of the door assembly quality. The operation curve characteristic database includes the operation curve characteristic data of qualified assembled doors and the operation curve characteristic data of doors in unqualified assembly states; After extracting the operation curve characteristics of the door after installation and adjustment, combine the operation curve characteristic data of qualified assembled doors and complete the determination of whether the assembly quality of the door is qualified through cluster analysis; when it is determined that the assembly quality of the door after installation and adjustment is unqualified, match the operation curve characteristics of the door with the operation curve characteristic data of doors in unqualified assembly states in the curve characteristic database to complete the determination of the unqualified state category of the door; When the door assembly is in a qualified state, calculate the assembly quality score of the door according to the deviation amount of the door operation curve characteristic value relative to the standard operation curve characteristic value.

2. The method for evaluating the installation and adjustment quality of rail transit doors according to claim 1, characterized in that: The relevant data of the unqualified door assembly output in step (2) is the unqualified state category of the door and / or the name of the unqualified door assembly part.

3. A system for evaluating the installation and adjustment quality of rail transit doors, characterized in that: It includes a door parameter analysis device with edge computing function and a data acquisition unit. The data acquisition unit is connected to the door parameter analysis device, and the door parameter analysis device is used to execute the method described in any one of claims 1-2.

4. The system for evaluating the installation and adjustment quality of rail transit doors according to claim 3, characterized in that: The door parameter analysis device is used to output the evaluation result of the door assembly quality. The door parameter analysis device is connected to a human-machine interaction screen, and the human-machine interaction screen is used to display the evaluation result of the door assembly quality. The door parameter analysis device is connected to an external terminal through a communication unit.

5. A rail vehicle, characterized in that: It includes a door controller and a door parameter analysis device with edge computing function. The door controller is connected to the door parameter analysis device, and the door parameter analysis device is used to execute the method described in any one of claims 1-2.

6. The rail vehicle according to claim 5, characterized in that: The door controller includes a main door controller and several slave door controllers for collecting the operation state data of individual doors. The main door controller and the slave door controllers form a communication ring network, and the main door controller is connected to the door parameter analysis device.

Images