Turnout state monitoring and fault prediction method and system based on multi-source curve

Abstract

The invention relates to the technical field of railway equipment monitoring, in particular to a turnout state monitoring and fault prediction method and system based on a multi-source curve, and the method comprises the steps: collecting current data during the working period of a turnout, generating an action current curve, and calibrating the time domain characteristics of a plurality of working subintervals of the turnout; monitoring the electrical state of the turnout, determining a work starting node of the turnout, and collecting an action image set of the turnout and a motor output detection data set of the turnout in combination with the time domain features; and according to the action image set and the motor output detection data set, generating an action parameter curve and outputting the action parameter curve so as to determine an action switching abnormal event of the turnout, so that the fault state of the turnout is predicted, and an early warning notification is generated. All working subintervals of the turnout are delimited by constructing an action current curve, so that an action parameter curve of the turnout and an output action parameter curve of a motor of the turnout are constructed, various curves are used for representing hardware abnormity of the turnout, potential faults are accurately predicted, and the safety of the turnout is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of railway equipment monitoring, in particular to a turnout state monitoring and fault prediction method and system based on multi-source curves. BACKGROUND

[0002] The turnout is an important railway signal device, mainly composed of a switch machine (including a point rail), a connecting rail, a frog, and a guard rail. The point rail guides the train wheels to turn by swinging left and right. The efficient and accurate operation of the turnout directly affects the train operation efficiency and safety. The existing technology represents the working current change of the turnout during the entire action period by constructing an action current curve. Considering that the above working current corresponds to the working current of the motor driving the turnout, the above action current curve can only reflect the response state and driving output state of the motor itself, such as whether the motor has a response delay and the driving output torque of the motor, etc. It cannot truly reflect the mechanical hardware state of the turnout itself. The hardware components in the turnout, especially the point rail, will wear out during long-term repeated action. The action current curve cannot accurately identify the hardware failure of the turnout, reducing the reliability and accuracy of the turnout monitoring. Therefore, how to dynamically monitor and predict the fault of the turnout at the hardware level is of great significance to improve the electrical and mechanical safety of the turnout. SUMMARY

[0003] Considering that the existing technology relies on a single action current curve to only monitor the performance of the turnout at the electrical level, the hardware failure of the internal components of the turnout formed during long-term repeated switching cannot be accurately detected in time, reducing the reliability and accuracy of the turnout monitoring. In view of the above problems, the present application is proposed to provide a turnout state monitoring and fault prediction method based on multi-source curves to overcome the above problems or at least partially solve the above problems, comprising: Step S1: collecting current data during the operation of the turnout to generate an action current curve; and calibrating the time domain features of a plurality of working subintervals of the turnout from the action current curve; Step S2: monitoring the electrical state of the turnout to determine the working start node of the turnout; and collecting the action image set of the turnout itself and the motor output detection data set of the turnout according to the working start node and the time domain features; Step S3: generating an action parameter curve and an output action parameter curve according to the action image set and the motor output detection data set; and determining the action switching abnormal event of the turnout by comparing the action parameter curves and the output action parameter curves of the turnout during a plurality of working periods; Step S4: predicting the fault state of the turnout according to the action switching abnormal event to generate a warning notification.

[0004] Optionally, in step S1, current data during the turnout operation is collected to generate an action current curve; time domain features of several operation subintervals of the turnout are calibrated from the action current curve, including: The action state of the starting electrical equipment of the turnout is monitored to determine whether the turnout enters an operation cycle; The respective full-process current data of the turnout during several operation cycles is collected, the interference intensity time domain distribution of the full-process current data of each operation cycle is identified, and the credibility of the full-process current data of each operation cycle is determined; The credibility of all full-process current data is compared to generate an effective action current curve of the turnout; The time domain variation of the effective action current curve is identified to determine all turning points of the effective action current curve; according to the all turning points, the time range features covered by the respective several operation subintervals of the turnout are calibrated; the several operation subintervals respectively correspond to the several hardware action stages of the turnout.

[0005] Optionally, in step S2, the electrical state of the turnout is monitored to determine the operation starting node of the turnout; according to the operation starting node and the time domain features, the action image set of the turnout itself and the motor output detection data set of the turnout are collected, including: The trigger instruction received by the starting electrical equipment of the turnout is monitored, and the time point at which the trigger instruction is successfully executed is taken as the operation starting node of the turnout; The starting time point of the operation starting node is taken as the starting time point, and according to the time ranges corresponding to the several operation subintervals contained in the time domain features, the shooting parameter time domain variation strategy for the actual scene where the turnout is located and the motor output detection time domain variation strategy for the turnout are adjusted; According to the shooting parameter time domain variation strategy and the motor output detection time domain variation strategy, the action image set of the turnout itself and the motor output detection data set of the turnout are collected; the action image set includes several turnout entity action image subsets corresponding to the several operation subintervals, and the motor output detection data set includes several motor output torque data subsets corresponding to the several operation subintervals.

[0006] Optionally, in step S3, according to the action image set and the motor output detection data set, an action parameter curve and an output action parameter curve are generated; the action parameter curve and the output action parameter curve of the turnout during several operations are compared to determine the action switching abnormal event of the turnout, including: Identify all the switch entity action image subsets under the action image set, obtain the action trajectory of the switch entity corresponding to the several work subintervals, and integrate to generate an action parameter curve; Integrate the motor output torque data subset under the motor output detection data set according to time sequence, and generate an output action parameter curve; Compare the action parameter curve and the output action parameter curve of the switch during the same work period, and obtain the trajectory deviation between the switch action trajectory and the switch actual action trajectory expected to be generated by the motor output torque of the switch during the same work period; Compare the trajectory deviations of the switch during several work periods according to time lapse, determine the trajectory deviation accumulation of the switch during several hardware action stages, and thus determine the action switching abnormal event of the switch.

[0007] Optionally, in step S4, according to the action switching abnormal event, the fault state of the switch is predicted, and thus a warning notification is generated, including: According to the hardware action stage of the switch corresponding to the action switching abnormal event, the potential wear parts of the switch are marked; According to the action trajectory deviation trend of the potential wear parts during the several work periods, the time information of the failure of the switch is predicted, and thus a warning notification is generated.

[0008] As an aspect of the present application, the embodiments of the present application also provide a switch state monitoring and fault prediction system based on multiple source curves, including: A first curve generation module is configured to collect current data during switch work, and generate an action current curve; A work interval marking module is configured to mark the time domain features of several work subintervals of the switch from the action current curve; A switch start determination module is configured to monitor the electrical state of the switch, and determine the work start node of the switch; A collection module is configured to collect the action image set of the switch itself and the motor output detection data set of the switch according to the work start node and the time domain features; A second curve generation module is configured to generate an action parameter curve and an output action parameter curve according to the action image set and the motor output detection data set; An abnormal event determination module is configured to compare the action parameter curves and the output action parameter curves of the switch during several work periods, and determine the action switching abnormal event of the switch; A fault prediction and warning module is configured to predict the fault state of the switch according to the action switching abnormal event, and thus generate a warning notification.

[0009] Optionally, the first curve generation module is configured to collect current data during switching operation of the turnout to generate a motion current curve, including: monitoring a motion state of a starting electrical device of the turnout to determine whether the turnout enters an operation period; collecting respective full-process current data of the turnout during a plurality of operation periods, performing interference intensity time domain distribution identification on the full-process current data of each operation period, and determining a credibility of the full-process current data of each operation period; comparing the credibility of all the full-process current data to generate an effective motion current curve of the turnout; The working interval calibration module is configured to calibrate time domain features of a plurality of working subintervals of the turnout from the motion current curve, including: performing time domain change identification on the effective motion current curve to determine all turning points of the effective motion current curve, and calibrating time range features covered by the plurality of working subintervals respectively according to the all turning points, wherein the plurality of working subintervals respectively correspond to a plurality of hardware motion stages of the turnout.

[0010] Optionally, the turnout starting determination module is configured to monitor an electrical state of the turnout to determine a working starting node of the turnout, including: monitoring a trigger instruction received by a starting electrical device of the turnout, and taking a time point at which the trigger instruction is successfully executed as the working starting node of the turnout; The collection module is configured to collect a motion image set of the turnout itself and a motor output detection data set of the turnout according to the working starting node and the time domain features, including: taking the working starting node as a starting time point, and adjusting a shooting parameter time domain change strategy for an actual scene in which the turnout is located and a motor output detection time domain change strategy for the turnout according to a plurality of time ranges corresponding to the plurality of working subintervals included in the time domain features; collecting the motion image set of the turnout itself and the motor output detection data set of the turnout according to the shooting parameter time domain change strategy and the motor output detection time domain change strategy; wherein the motion image set includes a plurality of turnout entity motion image subsets corresponding to the plurality of working subintervals, and the motor output detection data set includes a plurality of motor output torque data subsets corresponding to the plurality of working subintervals.

[0011] Optionally, the second curve generation module is configured to generate a motion parameter curve and an output action parameter curve according to the motion image set and the motor output detection data set, including: Identify all the switch entity action image subsets under the action image set, obtain the action trajectory of the switch entity corresponding to the several work subintervals, and integrate to generate an action parameter curve; According to the time sequence integration of the motor output torque data subset under the motor output detection data set, an output action parameter curve is generated. The abnormal event determination module is configured to compare the action parameter curve and the output action parameter curve of the switch during several work periods to determine the action switching abnormal event of the switch, including: Compare the action parameter curve and the output action parameter curve of the switch during the same work period to obtain the trajectory deviation between the expected switch action trajectory generated by the motor output torque of the switch during the same work period and the actual switch action trajectory. Compare the trajectory deviations of the switch during several work periods according to the time lapse to determine the trajectory deviation accumulation of the switch during several hardware action stages, thereby determining the action switching abnormal event of the switch.

[0012] Optionally, the fault prediction and early warning module is configured to predict the fault state of the switch according to the action switching abnormal event, and generate an early warning notification, including: According to the hardware action stage of the switch corresponding to the action switching abnormal event, the potential wear parts of the switch are marked; According to the action trajectory deviation trend of the potential wear parts during the several work periods, the time information of the failure of the switch is predicted, and an early warning notification is generated.

[0013] The beneficial effects of the above technical solutions provided in the embodiments of the present application at least include: The switch state monitoring and fault prediction method and system based on multiple source curves provided in the embodiments of the present application collect current data during the work of the switch to generate an action current curve, from which the time domain features of several work subintervals of the switch are marked. The electrical state of the switch is monitored to determine the work start node of the switch, and the action image set of the switch itself and the motor output detection data set of the switch are collected in combination with the time domain features. According to the action image set and the motor output detection data set, the action parameter curve and the output action parameter curve are generated to determine the action switching abnormal event of the switch, thereby predicting the fault state of the switch and generating an early warning notification. By constructing the action current curve to demarcate all work subintervals of the switch, the action parameter curve of the switch and the output action parameter curve of the motor are constructed, and the hardware abnormalities of the switch are represented by multiple curves to accurately predict potential faults and improve the electrical and mechanical safety of the switch.

[0014] Other features and advantages of the present application will be set forth in the description that follows, and in part will be apparent from the description, or can be learned by practice of the application. The purposes and other advantages of the present application will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.

[0015] The technical solutions of the present application will be further described in detail below with the aid of the accompanying drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0016] The accompanying drawings are included to provide a further understanding of the present application and are incorporated in and constitute a part of the specification, illustrate embodiments of the present application and are used to explain the present application, but do not constitute a limitation on the present application. In the drawings: Figure 1 A flowchart of the method for monitoring and predicting the turnout state based on multi-source curves provided in the embodiments of the present application is shown in the figure. Figure 2 A structural diagram of the system for monitoring and predicting the turnout state based on multi-source curves provided in the embodiments of the present application is shown in the figure. DETAILED DESCRIPTION

[0017] Exemplary embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. Rather, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be accurately conveyed to those skilled in the art.

[0018] In the description of the present application, it should be noted that the terms "upper", "lower", "left", "right", "rear" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application. In addition, the terms "first", "second", "third" are only for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0019] In the description of the present application, it should be noted that unless otherwise explicitly specified and limited, the term "connection" should be understood broadly, for example, it can be a fixed connection, or a detachable connection, or an integral connection; it can be a mechanical connection, or an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, or it can be a connection between two elements. For those skilled in the art, the specific meaning of the above-mentioned term in the present application can be understood according to the specific circumstances.

[0020] Reference should be made to Figure 1As shown, an embodiment of the present application provides a turnout state monitoring and fault prediction method based on multi-source curves. The turnout state monitoring and fault prediction method based on multi-source curves comprises: Step S1: collecting current data during turnout operation to generate an action current curve; and determining time domain features of a plurality of operation subintervals of the turnout from the action current curve; Step S2: monitoring an electrical state of the turnout to determine an operation start node of the turnout; collecting an action image set of the turnout and a motor output detection data set of the turnout according to the operation start node and the time domain features; Step S3: generating an action parameter curve and an output action parameter curve according to the action image set and the motor output detection data set; and determining an action switching abnormal event of the turnout by comparing the action parameter curve and the output action parameter curve of the turnout during a plurality of operation periods; Step S4: predicting a fault state of the turnout according to the action switching abnormal event to generate an early warning notification.

[0021] The turnout state monitoring and fault prediction method based on multi-source curves can accurately predict potential faults and improve electrical and mechanical safety of the turnout by constructing an action current curve to demarcate all operation subintervals of the turnout, constructing an action parameter curve of the turnout and an output action parameter curve of a motor thereof, and using a plurality of curves to represent hardware abnormalities of the turnout.

[0022] In another embodiment, in step S1, current data during operation of the turnout is collected to generate an action current curve; and time domain features of a plurality of operation subintervals of the turnout are determined from the action current curve, comprising: monitoring an action state of a start electrical device of the turnout to determine whether the turnout enters an operation period; collecting whole-process current data of the turnout during a plurality of operation periods; performing interference intensity time domain distribution identification on the whole-process current data of each operation period to determine a credibility of the whole-process current data of each operation period; comparing the credibility of all whole-process current data to generate an effective action current curve of the turnout; performing time domain change identification on the effective action current curve to determine all turning points of the effective action current curve; determining time range features covered by each of a plurality of operation subintervals of the turnout according to the all turning points; and the plurality of operation subintervals respectively correspond to a plurality of hardware action stages of the turnout.

[0023] In the above technical solution, the main components of the turnout include the switch (including the point), the connecting rail, the frog, and the guard rail. In actual operation, a motor such as a three-phase AC motor can be used to drive the turnout action, especially the action of the point of the switch. The motor provides a matching driving action to the switch according to the size and direction of the wheel turning required by the turnout, and the above driving action depends on the size of the motor. In addition, the turnout is triggered to work when the wheel turning needs to be performed, and the turnout is in a non-working state at other times, that is, the motor is only applied with a corresponding driving current when the turnout is triggered to work, and the size of the driving current applied to the motor directly reflects the actual working state of the turnout, that is, when the driving current is in a corresponding size range, the turnout is in a corresponding hardware action stage (i.e., the point action stage).

[0024] In order to determine the time range corresponding to all hardware action stages of the turnout in the entire working cycle, the full-process current data of the turnout in several working cycles (i.e., the working cycle corresponding to the working state of the turnout after being triggered to start each time) is collected first. The full-process current data corresponding to each working cycle reflects the current size applied to the turnout by the motor. The motor driving the working process of the turnout is affected by internal and external factors, and the collected full-process current data has noise interference and covers the size change of the current itself due to driving demand. In order to accurately divide all working subintervals of the turnout in a complete working cycle, the disturbance intensity time domain distribution of the full-process current data of each working cycle is identified, that is, the average interference noise intensity of the full-process current data in the entire time range is determined, so as to determine the credibility of the full-process current data. It can be understood that the smaller the average interference noise intensity, the higher the credibility. By comparing the credibility of all full-process current data, the full-process current data with the smallest credibility is taken as the original data, and the effective action current curve of the turnout is generated, that is, the above effective action current curve reflects the change of the current size with time in the full-process current data with the smallest credibility.

[0025] The motor drives the turnout (especially the point rail) to work, which is a dynamic process. The point rail moves from the initial position to the target position, including multiple action stages, such as unlocking stage, conversion stage, locking stage, and slow release stage. In each of the above stages, the driving current applied by the motor is different. By identifying the time domain changes of the effective action current curve and determining all turning points of the effective action current curve, the above turning points can be, but are not limited to, the curve inflection points of the effective action current curve. Then, taking all turning points as the reference, the time range characteristics covered by each of the several working subintervals of the turnout on the time axis of the above effective action current curve are divided. It can be understood that each working subinterval corresponds to a hardware action stage, that is, all working subintervals one by one correspond to the above unlocking stage, conversion stage, locking stage, and slow release stage. By defining the time range characteristics covered by each of the working subintervals, a time reference is provided for subsequent time domain change analysis and comparison of the action parameter curve and the output action parameter curve.

[0026] In another embodiment, in step S2, the electrical state of the turnout is monitored to determine the working start node of the turnout; and according to the working start node and the time domain characteristics, the action image set of the turnout itself and the motor output detection data set of the turnout are collected, including: Monitoring the trigger instruction received by the starting electrical equipment of the turnout, and taking the time point at which the trigger instruction is successfully executed as the working start node of the turnout; Taking the working start node as the starting time point, adjusting the shooting parameter time domain change strategy for the actual scene of the turnout and the motor output detection time domain change strategy for the turnout according to the time ranges of the one-to-one corresponding several working subintervals contained in the time domain characteristics; According to the shooting parameter time domain change strategy and the motor output detection time domain change strategy, the action image set of the turnout itself and the motor output detection data set of the turnout are collected; wherein the action image set includes several turnout entity action image subsets corresponding to the several working subintervals, and the motor output detection data set includes several motor output torque data subsets corresponding to the several working subintervals.

[0027] In the above technical solution, the action current curve only reflects the working performance of the turnout and the motor at the electrical level, but the turnout as a hardware device inevitably wears out during long-term repeated action, resulting in response delay and failure to move to the target position of the turnout under the same motor driving effect, reducing the working reliability and accuracy of the turnout. In order to identify the abnormal situation of the turnout at the mechanical hardware level, the trigger instruction (such as the electrical pulse instruction received by the relay device) received by the starting electrical equipment (such as the relay device) of the turnout is monitored during the actual working process of the turnout, and the time point at which the starting electrical equipment successfully performs the corresponding engagement action after responding to the trigger instruction is taken as the working start (time) node of the turnout.

[0028] When the turnout successfully acts, the above-mentioned work starting (time) node is taken as a starting time point, and the shooting parameter time domain change strategy for the actual scene where the turnout is located and the motor output detection time domain change strategy for the turnout are adjusted in combination with the time ranges covered by the respective work subintervals; it can be understood that the shooting parameter time domain change strategy refers to the change strategy of the actual shooting parameters (such as shooting focal length and / or shooting field angle, etc.) for the turnout action in the above-mentioned time ranges starting from the above-mentioned starting time point; the motor output detection time domain change strategy refers to the detection parameters (such as detection sensitivity and / or detection sampling frequency, etc.) for the motor output (torque) detection in the above-mentioned time ranges starting from the above-mentioned starting time point. Then, according to the shooting parameter time domain change strategy and the motor output detection time domain change strategy, the action image set of the turnout itself and the motor output detection data set of the turnout are respectively collected, so as to obtain the action image subsets of the turnout entity corresponding to the work subintervals and the motor output torque data subsets, which provide sufficient data for subsequent identification of the hardware wear state of the turnout.

[0029] In another embodiment, in step S3, the action parameter curve and the output action parameter curve are generated according to the action image set and the motor output detection data set; by comparing the action parameter curves and the output action parameter curves of the turnout during the work periods, the action switching abnormal event of the turnout is determined, including: The action image subsets belonging to the action image set of the turnout are identified to obtain the action trajectories of the turnout entity corresponding to the work subintervals, so as to generate the action parameter curve; The motor output torque data subsets belonging to the motor output detection data set are integrated in time sequence to generate the output action parameter curve; The action parameter curves and the output action parameter curves of the turnout during the same work period are compared to obtain the trajectory deviation between the action trajectory of the turnout expected to be generated by the motor output torque and the actual action trajectory of the turnout during the same work period; The trajectory deviations of the turnout during the work periods are compared according to time elapse to determine the trajectory deviation accumulation of the turnout during the hardware action stages, so as to determine the action switching abnormal event of the turnout.

[0030] In the technical solution, the action image set and the motor output detection data set correspond to the action state of the hardware components of the turnout (especially the point rail) in actual work and the torque state of the motor output to the hardware components. Therefore, the action image set is identified to obtain the action trajectory of the turnout entity corresponding to the working sub-interval, and the action trajectory is integrated to generate an action parameter curve (a continuous curve of the position of the action trajectory changing over time). The motor output torque data subset of the motor output detection data set is integrated in time sequence to generate an output action parameter curve (a continuous curve of the size of the output torque changing over time). It can be understood that, considering that the hardware wear of the turnout is a relatively long process, the curve analysis needs to be performed during multiple working periods of the turnout (during which the turnout performs multiple wheel steering tasks over time). Specifically, the trajectory deviation between the action trajectory of the turnout expected to be generated by the motor output torque and the actual action trajectory of the turnout during the same working period is obtained by comparing the action parameter curve and the output action parameter curve of the turnout during the same working period. The action trajectory of the turnout expected to be generated by the motor output torque can be determined according to the output action parameter curve, and the actual action trajectory of the turnout can be determined according to the action parameter curve. The trajectory deviation of the turnout during the working periods is compared over time to determine the trajectory deviation accumulation of the turnout during the hardware action stages. If the trajectory deviation accumulation exceeds the preset deviation threshold, it is determined that the action switching abnormal event of the turnout exists; otherwise, it is determined that the action switching abnormal event of the turnout does not exist. The fault state of the turnout can be accurately determined at the hardware level through the above method.

[0031] In another embodiment, in step S4, the fault state of the turnout is predicted according to the action switching abnormal event to generate a warning notification, including: According to the hardware action stage of the turnout corresponding to the action switching abnormal event, the potential wear component of the turnout is determined. According to the action trajectory deviation trend of the potential wear component during the working periods, the time information of the failure of the turnout is predicted to generate a warning notification.

[0032] In the technical solution, it can be understood that the unlocking stage, the conversion stage, the locking stage, and the buffer release stage of the turnout require the cooperation of different hardware components in the turnout. According to the hardware action stage of the turnout corresponding to the action switching abnormal event, the potential wear component of the turnout is determined to accurately determine the mechanical wear of the hardware components in the turnout. In addition, according to the action trajectory deviation trend of the potential wear component during the working periods, the occurrence time of the action trajectory deviation of the potential wear component reaching the acceptable deviation limit value is estimated, and the occurrence time is taken as the time information of the failure of the turnout. Therefore, a warning notification is generated and sent to the platform to realize the safety monitoring of the turnout.

[0033] Referring to Figure 2 An embodiment of the present application provides a turnout state monitoring and fault prediction system based on multi-source curves. The turnout state monitoring and fault prediction system based on multi-source curves comprises: A first curve generation module is configured to collect current data during turnout operation to generate a movement current curve; A working interval calibration module is configured to calibrate time domain features of a plurality of working subintervals of the turnout from the movement current curve; A turnout start determination module is configured to monitor an electrical state of the turnout and determine a working start node of the turnout; A collection module is configured to collect a movement image set of the turnout and a motor output detection data set of the turnout according to the working start node and the time domain features; A second curve generation module is configured to generate a movement parameter curve and an output action parameter curve according to the movement image set and the motor output detection data set; An abnormal event determination module is configured to compare the movement parameter curve and the output action parameter curve of the turnout during a plurality of working periods to determine a movement switching abnormal event of the turnout; A fault prediction and early warning module is configured to predict a fault state of the turnout according to the movement switching abnormal event to generate an early warning notification.

[0034] The turnout state monitoring and fault prediction system based on multi-source curves can accurately predict potential faults and improve the electrical and mechanical safety of the turnout by constructing a movement current curve to demarcate all working subintervals of the turnout, constructing a movement parameter curve of the turnout and an output action parameter curve of the motor thereof, and using a plurality of curves to represent hardware abnormalities of the turnout.

[0035] In another embodiment, the first curve generation module is configured to collect current data during turnout operation to generate a movement current curve, and comprises: The movement state of a start electrical device of the turnout is monitored to determine whether the turnout enters a working period; Full-process current data of the turnout during a plurality of working periods are collected, the full-process current data of each working period are subjected to interference intensity time domain distribution identification, and the reliability of the full-process current data of each working period is determined; The reliabilities of all full-process current data are compared to generate an effective movement current curve of the turnout; The working interval calibration module is configured to calibrate time domain features of a plurality of working subintervals of the turnout from the movement current curve, and comprises: The time domain change recognition is performed on the effective action current curve to determine all turning change points of the effective action current curve; and the time range characteristics covered by the respective working subintervals of the switch are calibrated according to all the turning change points; wherein the working subintervals one by one correspond to the hardware action stages of the switch.

[0036] In another embodiment, the switch starting determination module is configured to monitor the electrical state of the switch, and determine the working starting node of the switch, including: monitoring the trigger instruction received by the starting electrical equipment of the switch, and taking the time point at which the trigger instruction is successfully executed as the working starting node of the switch; The acquisition module is configured to acquire the action image set of the switch itself and the motor output detection data set of the switch according to the working starting node and the time domain characteristics, including: taking the working starting node as the starting time point, and adjusting the shooting parameter time domain change strategy for the actual scene in which the switch is located and the motor output detection time domain change strategy for the switch according to the time ranges of the one-to-one corresponding working subintervals included in the time domain characteristics; acquiring the action image set of the switch itself and the motor output detection data set of the switch according to the shooting parameter time domain change strategy and the motor output detection time domain change strategy; wherein the action image set includes the sub-set of the switch entity action images corresponding to the working subintervals one by one, and the motor output detection data set includes the sub-set of the motor output torque data corresponding to the working subintervals one by one.

[0037] In another embodiment, the second curve generation module is configured to generate the action parameter curve and the output action parameter curve according to the action image set and the motor output detection data set, including: identifying all the sub-sets of the switch entity action images under the action image set to obtain the action trajectories of the switch entity corresponding to the working subintervals, and integrating the action parameter curve therefrom; integrating the sub-sets of the motor output torque data according to the time sequence to generate the output action parameter curve; The abnormal event determination module is configured to compare the action parameter curve and the output action parameter curve of the switch during the working periods to determine the action switching abnormal event of the switch, including: comparing the action parameter curve and the output action parameter curve of the switch during the same working period to obtain the trajectory deviation between the expected switch action trajectory of the motor output torque of the switch during the same working period and the actual switch action trajectory of the switch; comparing the trajectory deviations of the switch during the working periods according to the time elapse to determine the trajectory deviation accumulation of the switch during the hardware action stages, and thereby determine the action switching abnormal event of the switch.

[0038] In another embodiment, a fault prediction and early warning module is used to predict the fault state of the turnout according to the action switching abnormal event, and to generate a warning notification, including: According to the hardware action stage of the action switching abnormal event corresponding to the turnout, the potential wear parts of the turnout are determined; According to the action trajectory deviation trend of the potential wear parts during several work periods, the time information of the turnout failure is predicted, and a warning notification is generated.

[0039] The operation and effect of the turnout state monitoring and fault prediction system based on multi-source curves of the present application are consistent with the operation and effect of the above-mentioned turnout state monitoring and fault prediction method based on multi-source curves. Therefore, the turnout state monitoring and fault prediction system based on multi-source curves will not be repeated here.

[0040] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. The present disclosure is not limited to the precise structure described above and shown in the drawings, and various modifications and changes can be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application also intends to include these modifications and variations.

Claims

1. A method for monitoring and failure prediction of turnout state based on multi-source curve, characterized in that, Comprising: Step S1: collecting current data during the operation of the turnout to generate an action current curve; and calibrating time domain features of a plurality of operation subintervals of the turnout from the action current curve; Step S2: monitoring the electrical state of the turnout to determine an operation start node of the turnout; and collecting an action image set of the turnout itself and a motor output detection data set of the turnout according to the operation start node and the time domain features; Step S3: generating an action parameter curve and an output action parameter curve according to the action image set and the motor output detection data set; and determining an action switching abnormal event of the turnout by comparing the action parameter curve and the output action parameter curve of the turnout during a plurality of operations; Step S4: predicting a fault state of the turnout according to the action switching abnormal event to generate a warning notification.

2. The turnout state monitoring and fault prediction method based on multiple source curves according to claim 1, wherein: in step S1, the current data during the operation of the turnout is collected to generate an action current curve; and the time domain features of a plurality of operation subintervals of the turnout are calibrated from the action current curve, comprising: monitoring the action state of the starting electrical equipment of the turnout to determine whether the turnout enters an operation cycle; collecting the full-process current data of the turnout during a plurality of operation cycles; performing interference intensity time domain distribution identification on the full-process current data of each operation cycle to determine the credibility of the full-process current data of each operation cycle; comparing the credibility of all full-process current data to generate an effective action current curve of the turnout; performing time domain change identification on the effective action current curve to determine all turning points of the effective action current curve; and calibrating time range features covered by a plurality of operation subintervals according to the all turning points; wherein the plurality of operation subintervals correspond to a plurality of hardware action stages of the turnout one by one.

3. The turnout state monitoring and fault prediction method based on multiple source curves according to claim 1, wherein: in step S2, the electrical state of the turnout is monitored to determine an operation start node of the turnout; and the action image set of the turnout itself and the motor output detection data set of the turnout are collected according to the operation start node and the time domain features, comprising: monitoring the trigger instruction received by the starting electrical equipment of the turnout to take the time point at which the trigger instruction is successfully executed as the operation start node of the turnout; adjusting the shooting parameter time domain change strategy for the actual scene where the turnout is located and the motor output detection time domain change strategy for the turnout according to a plurality of time ranges corresponding to the plurality of operation subintervals included in the time domain features, with the operation start node as the starting time point. ​ ​ ​ According to the shooting parameter time domain change strategy and the motor output detection time domain change strategy, the action image set of the turnout itself and the motor output detection data set of the turnout are collected; wherein the action image set includes a plurality of turnout entity action image subsets corresponding to the plurality of working subintervals, and the motor output detection data set includes a plurality of motor output torque data subsets corresponding to the plurality of working subintervals.

4. The turnout state monitoring and fault prediction method based on multi-source curves according to claim 1, wherein: In step S3, the action parameter curve and the output action parameter curve are generated according to the action image set and the motor output detection data set; By comparing the action parameter curves and the output action parameter curves of the turnout in a plurality of working periods, the action switching abnormal event of the turnout is determined, including: All the turnout entity action image subsets under the action image set are identified to obtain a plurality of action trajectories of the turnout corresponding to the plurality of working subintervals, so as to integrate and generate the action parameter curve; The motor output torque data subsets under the motor output detection data set are integrated according to time sequence to generate the output action parameter curve; By comparing the action parameter curve and the output action parameter curve of the turnout in the same working period, the trajectory deviation between the expected turnout action trajectory and the actual turnout action trajectory of the motor output torque of the turnout in the same working period is obtained; By comparing the trajectory deviations of the turnout in a plurality of working periods according to time lapse, the trajectory deviation accumulation of the turnout in a plurality of hardware action stages is determined, so as to determine the action switching abnormal event of the turnout.

5. The turnout state monitoring and fault prediction method based on multi-source curves according to claim 1, wherein: In step S4, the fault state of the turnout is predicted according to the action switching abnormal event, so as to generate a warning notification, including: According to the hardware action stage of the turnout corresponding to the action switching abnormal event, the potential wear part of the turnout is marked; According to the action trajectory deviation trend of the potential wear part in the plurality of working periods, the time information of the failure of the turnout is predicted, so as to generate a warning notification.

6. A system for monitoring and predicting failure of a turnout based on multi-source curves, characterized in that, Including: A first curve generation module for collecting current data during the working period of the turnout to generate an action current curve; A working interval marking module for marking the time domain characteristics of a plurality of working subintervals of the turnout from the action current curve; A turnout start determination module for monitoring the electrical state of the turnout to determine the working start node of the turnout; A collection module for collecting the action image set of the turnout itself and the motor output detection data set of the turnout according to the working start node and the time domain characteristics; A second curve generation module for generating an action parameter curve and an output action parameter curve according to the action image set and the motor output detection data set; An abnormal event determination module for comparing the action parameter curves and the output action parameter curves of the turnout in a plurality of working periods to determine the action switching abnormal event of the turnout; A fault prediction and early warning module is configured to predict a fault state of the turnout according to the action switching abnormal event, and generate a warning notification. 7.The turnout state monitoring and fault prediction system based on multi-source curves according to claim 6, wherein: The first curve generation module is configured to collect current data during turnout operation to generate an action current curve, including: monitoring an action state of a starting electrical device of the turnout to determine whether the turnout enters an operation cycle; collecting respective full-process current data of the turnout during a plurality of operation cycles, performing interference intensity time domain distribution identification on the full-process current data of each operation cycle, and determining a credibility of the full-process current data of each operation cycle; comparing the credibilities of all full-process current data to generate an effective action current curve of the turnout; The working interval calibration module is configured to calibrate time domain features of a plurality of working subintervals of the turnout from the action current curve, including: performing time domain change identification on the effective action current curve to determine all turning points of the effective action current curve, and calibrating time range features covered by the plurality of working subintervals according to the all turning points, wherein the plurality of working subintervals respectively correspond to a plurality of hardware action stages of the turnout. 8.The turnout state monitoring and fault prediction system based on multi-source curves according to claim 6, wherein: The turnout starting determination module is configured to monitor an electrical state of the turnout to determine a working starting node of the turnout, including: monitoring a trigger instruction received by a starting electrical device of the turnout, and taking a time point at which the trigger instruction is successfully executed as the working starting node of the turnout; The collection module is configured to collect an action image set of the turnout and a motor output detection data set of the turnout according to the working starting node and the time domain features, including: adjusting a shooting parameter time domain change strategy for an actual scene in which the turnout is located and a motor output detection time domain change strategy according to the time domain features including a plurality of time ranges corresponding to the plurality of working subintervals, with the working starting node as a starting time point; collecting the action image set of the turnout and the motor output detection data set of the turnout according to the shooting parameter time domain change strategy and the motor output detection time domain change strategy; wherein the action image set includes a plurality of turnout entity action image subsets corresponding to the plurality of working subintervals, and the motor output detection data set includes a plurality of motor output torque data subsets corresponding to the plurality of working subintervals. 9.The turnout state monitoring and fault prediction system based on multi-source curves according to claim 6, wherein: The second curve generation module is configured to generate an action parameter curve and an output action parameter curve according to the action image set and the motor output detection data set, including: performing identification on all turnout entity action image subsets under the action image set to obtain a plurality of action trajectories corresponding to the plurality of working subintervals of the turnout, and integrating to generate the action parameter curve; The motor output torque data subset belonging to the motor output detection data set is integrated in time sequence to generate an output action parameter curve; The abnormal event determination module is configured to compare the action parameter curve and the output action parameter curve of the turnout during several work periods to determine the action switching abnormal event of the turnout, including: The action parameter curve and the output action parameter curve of the turnout during the same work period are compared to obtain a track deviation between a track trajectory expected to be generated by the motor output torque of the turnout during the same work period and an actual track trajectory of the turnout; The track deviations of the turnout during several work periods are compared in time to determine a track offset accumulation of the turnout during several hardware action stages, thereby determining the action switching abnormal event of the turnout.

10. The turnout state monitoring and fault prediction system based on multi-source curves according to claim 6, wherein: The fault prediction and early warning module is configured to predict a fault state of the turnout according to the action switching abnormal event, and generate an early warning notification, including: According to the hardware action stage of the turnout corresponding to the action switching abnormal event, a potential wear component of the turnout is marked; According to the action track offset trend of the potential wear component during the several work periods, time information of a fault failure of the turnout is predicted, and an early warning notification is generated.

Images