Fault sign diagnostic device and fault sign diagnostic method

Abstract

To diagnose a sign of a failure in an air compressor or a relevant instrument thereof regardless of operation and linear conditions of a vehicle.SOLUTION: A failure sign diagnostic device 100 comprises: a data input unit 110 which inputs operation data of an air compressor 210 and pressure data of a relevant instrument; an operation time calculation unit 120 which calculates an actual operation time of the air compressor 210 on the basis of the operation data; an integration pressure displacement amount calculation unit 130 which calculates an integration pressure displacement amount indicating an amount of the compressed air used by the relevant instrument with respect to the operation of the air compressor 210 for each relevant instrument on the basis of the operation data and pressure data; an operation time prediction model unit 140 which calculates a predicted operation time of the air compressor 210 from a prediction model using a feature amount which is the integration pressure displacement amount of each relevant instrument; and a failure sign diagnostic unit 150 which diagnoses the presence / absence of a failure sign in at least one of the air compressor 210 and the relevant instrument on the basis of a comparison result between the actual operation time and the predicted operation time.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a fault prediction diagnosis device and a fault prediction diagnosis method, and is suitable for application to a fault prediction diagnosis device and a fault prediction diagnosis method for diagnosing a fault prediction in a pneumatic compressor used in a railway vehicle or related equipment using its compressed air.

Background Art

[0002] From the viewpoints of improving the efficiency and labor saving of maintenance work, the development of a fault prediction diagnosis technology that utilizes data acquired from on-vehicle equipment has been demanded. Among on-vehicle equipment, a pneumatic compressor that generates compressed air used in air brakes, air springs, etc. affects the operation of other equipment such as air brakes in case of a failure. Therefore, the development of a technology for detecting and diagnosing the prediction of a failure before it occurs is strongly demanded.

[0003] For example, Patent Document 1 discloses an abnormality detection method that attempts to detect the occurrence of an abnormality in a compressor with high accuracy using equipment monitoring data of a railway vehicle. According to the abnormality detection method of Patent Document 1, data when the compressor (pneumatic compressor) is operating and the brake is not operating is extracted, and further, among the data obtained by dividing the data at regular intervals, the data with the largest increase in the pressure of the accumulator system is extracted. Also, the amount of pressure increase is calculated by multiple regression analysis using, as explanatory variables, the amount of pressure displacement of the air spring at that time, the occupancy rate calculated from the vehicle weight, the traveling speed, and the outside air temperature information. Then, abnormality detection is performed by comparing the extracted increase in the pressure of the accumulator system with the increase in the pressure calculated by multiple regression analysis.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, in the abnormality detection method of Patent Document 1, data on the amount of pressure increase used for abnormality detection judgment is extracted for the time when the air compressor is operating and the equipment using compressed air (e.g., brakes) is not operating. Therefore, depending on the operation of the vehicle equipped with the equipment and the line conditions, the period corresponding to the data extraction conditions may be short, and there is a risk that sufficient data for failure prediction diagnosis cannot be obtained. Specifically, for example, when the equipment is a brake of a railway vehicle, since the railway vehicle decelerates with the brake from before the stop station and stops at the station, when the distance between stations is short, data on the amount of pressure increase cannot be obtained. Also, when there is a steep downhill slope continuously between stations, since the brake operates continuously for a long time, data on the amount of pressure increase cannot be obtained. Thus, in the abnormality detection method of Patent Document 1, in a situation where the related equipment using compressed air is almost operating during the operation of the air compressor, sufficient data for abnormality detection judgment cannot be obtained, and as a result, there may be a period during which abnormality detection of the air compressor or its related equipment cannot be performed.

[0006] The present invention has been made in consideration of the above points, and intends to propose a failure prediction diagnosis device and a failure prediction diagnosis method capable of diagnosing the sign of failure in an air compressor or its related equipment using compressed air regardless of the operation of the vehicle and the line conditions.

Means for Solving the Problem

[0007] In order to solve such problems, the present invention provides a failure prediction diagnosis device for diagnosing a failure prediction in an air compressor or related equipment using compressed air generated by the air compressor, the failure prediction diagnosis device including: a data input unit for inputting operation data of the air compressor and pressure data of the related equipment; an operation time calculation unit for calculating an actual operation time of the air compressor based on the operation data input to the data input unit; an integrated pressure displacement amount calculation unit for calculating, for each related equipment, an integrated pressure displacement amount indicating an amount of compressed air used by the related equipment for the operation of the air compressor based on the operation data and the pressure data input to the data input unit; an operation time prediction model unit for calculating a predicted operation time of the air compressor from a prediction model using the integrated pressure displacement amount of each related equipment calculated by the integrated pressure displacement amount calculation unit as a feature amount; and a failure prediction diagnosis unit for diagnosing the presence or absence of a failure prediction in at least one of the air compressor or the related equipment based on a comparison result between the actual operation time calculated by the operation time calculation unit and the predicted operation time calculated by the operation time prediction model unit.

[0008] Also, in order to solve such problems, in the present invention, there is provided a failure prediction diagnosis method by a failure prediction diagnosis device that diagnoses a failure prediction in an air compressor or related equipment that uses compressed air generated by the air compressor. The failure prediction diagnosis method includes: a data input step in which the failure prediction diagnosis device inputs operation data of the air compressor and pressure data of the related equipment; an operation time calculation step in which the failure prediction diagnosis device calculates an actual operation time of the air compressor based on the operation data input in the data input step; an integrated pressure displacement amount calculation step in which the failure prediction diagnosis device calculates, for each related equipment, an integrated pressure displacement amount indicating the amount of compressed air used by the related equipment for the operation of the air compressor based on the operation data and the pressure data input in the data input step; an operation time prediction step in which the failure prediction diagnosis device calculates a predicted operation time of the air compressor from a prediction model that uses, as a feature amount, the integrated pressure displacement amount of each related equipment calculated in the integrated pressure displacement amount calculation step; and a failure prediction diagnosis step in which the failure prediction diagnosis device diagnoses the presence or absence of a failure prediction in at least one of the air compressor or the related equipment based on a comparison result between the actual operation time calculated in the operation time calculation step and the predicted operation time calculated in the operation time prediction step. Break A failure prediction diagnosis method including the above steps is provided.

Effects of the Invention

[0009] According to the present invention, it is possible to diagnose a failure prediction in an air compressor or its related equipment regardless of the operation and linear conditions of the vehicle.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Mode for Carrying Out the Invention

[0011] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[0012] (1) First Embodiment (1-1) Configuration FIG. 1 is a block diagram showing a configuration example of the failure omen diagnosis device 100 according to the first embodiment of the present invention. As shown in FIG. 1, the failure omen diagnosis device 100 includes a data input unit 110, an operation time calculation unit 120, an integrated pressure displacement amount calculation unit 130, an operation time prediction model unit 140, and a failure omen diagnosis unit 150.

[0013] In addition, an air compressor 210, a brake device 220, and an air spring 230 are connected to the fault prediction diagnosis device 100 as devices to be diagnosed for fault prediction. The air compressor 210 is a device that generates and supplies compressed air. The brake device 220 and the air spring 230 are examples of related devices that use the compressed air generated by the air compressor 210. The number of related devices of the air compressor 210 may be one or more.

[0014] The data input unit 110 inputs the operation data of the air compressor 210 from the air compressor 210, and inputs the respective pressure data from the brake device 220 and the air spring 230 that use the compressed air generated by the air compressor 210. Further, the data input unit 110 outputs the operation data of the air compressor 210 to the operation time calculation unit 120 and the integrated pressure displacement amount calculation unit 130, and outputs the pressure data of the brake device 220 and the air spring 230 to the integrated pressure displacement amount calculation unit 130.

[0015] In this embodiment, as the pressure data of the related devices that use compressed air, the brake cylinder (BC) pressure value of the brake device 220, which is often mounted in a general railway vehicle, and the air suspension (AS) pressure value of the air spring 230 will be used as examples for explanation. However, the pressure data that can be used in this embodiment is not limited to the above-described pressure data. When there are related devices that use the compressed air generated by the air compressor 210, the pressure data for each such related device may also be input to the data input unit 110. The same applies to other embodiments.

[0016] In addition, as for the method of inputting each data in the data input unit 110, data may be input by directly connecting the data input unit 110 to each device, or when a vehicle information control device (not shown) that aggregates and controls the information of each device is provided in a vehicle (for example, a railway vehicle) equipped with each device, each data may be input by connecting the data input unit 110 to the vehicle information control device. Further, when the vehicle equipped with each device has a function capable of transmitting each data to a storage area of a server installed at a vehicle base or the like or a server constructed on a cloud environment using public wireless communication or the like, the failure prediction diagnosis device 100 may be implemented in an environment where it can be connected to the above server, and each data may be input from the above server to the data input unit 110. The above also applies to other embodiments.

[0017] The operation time calculation unit 120 inputs the operation data of the air compressor 210 from the data input unit 110, calculates the actual operation time based on the input information, and outputs this to the failure prediction diagnosis unit 150. The operation data of the air compressor 210 indicates at least the operation state (operating or stopped) of the air compressor 210. Details of the processing by the operation time calculation unit 120 will be described later.

[0018] The integrated pressure displacement calculation unit 130 inputs the operation data, the BC pressure value, and the AS pressure value from the data input unit 110, and calculates the integrated pressure displacement (integrated BC pressure displacement, integrated AS pressure displacement) of each related device based on the input information, and outputs these to the operation time prediction model unit 140. The integrated pressure displacement is the displacement amount of the pressure of the related device with respect to one operation (which may be a predetermined number of times) of the air compressor 210, and is calculated for each related device. It should be noted that the integrated pressure displacement can also be said to be the amount of compressed air used by the related device during one operation of the air compressor 210. Details of the processing by the integrated pressure displacement calculation unit 130 will be described later.

[0019] The operating time prediction model unit 140 receives the integrated BC pressure displacement amount and the integrated AS pressure displacement amount from the integrated pressure displacement amount calculation unit 130, derives a predicted operating time that is a predicted value of the operating time based on the input information, and outputs this to the fault omen diagnosis unit 150. Details of the processing by the operating time prediction model unit 140 will be described later.

[0020] The fault omen diagnosis unit 150 receives the actual operating time from the operating time calculation unit 120 and the predicted operating time from the operating time prediction model unit 140, diagnoses the fault omen based on the input information, and outputs the diagnosis result to the driver or the vehicle base. Details of the processing by the fault omen diagnosis unit 150 will be described later.

[0021] Hereinafter, details of the processing by each functional unit of the fault omen diagnosis device 100 will be described.

[0022] (1-2) Processing of the operating time calculation unit 120 FIG. 2 is a flowchart showing an example of the processing procedure by the operating time calculation unit 120. The processing shown in FIG. 2 is repeatedly executed at a predetermined cycle.

[0023] According to FIG. 2, first, the operating time calculation unit 120 acquires the operation data of the air compressor 210 from the data input unit 110. As described above, the operation data of the air compressor 210 indicates at least the operating state (operating / stopped) of the air compressor 210 (step S101).

[0024] Next, the operating time calculation unit 120 refers to the operation data acquired in step S101 and determines whether the operating state of the air compressor 210 has changed from stopped to operating (step S102). If the operating state has changed from stopped to operating (YES in step S102), the process proceeds to step S103. If the operating state has not changed from stopped to operating (NO in step S102), the process proceeds to step S104.

[0025] In step S103, the operating time calculation unit 120 starts measuring the operating time and proceeds to step S104.

[0026] In step S104, the operation time calculation unit 120 refers to the operation data acquired in step S101 and determines whether the operation state of the air compressor 210 has changed from the operating state to the stopped state. If the operation state has changed from the operating state to the stopped state (YES in step S104), the process proceeds to step S105. If the operation state has not changed from the operating state to the stopped state (NO in step S104), the process ends.

[0027] In step S105, the operation time calculation unit 120 stops measuring the operation time, sets the measured time as the actual operation time, and proceeds to step S106. The actual operation time is the measured result of the operation time elapsed from when the air compressor 210 starts operating until it stops, and the operation time calculation unit 120 stores the actual operation time in a predetermined storage area.

[0028] Next, the operation time calculation unit 120 outputs the actual operation time set in step S105 to the failure prediction diagnosis unit 150 (step S106) and ends the process.

[0029] As described above, by executing the process shown in FIG. 2, the operation time calculation unit 120 can calculate and output the actual operation time of the air compressor 210 based on the operation data of the air compressor 210.

[0030] (1-3) Processing of the integrated pressure displacement amount calculation unit 130 FIG. 3 is a flowchart showing an example of the processing procedure by the integrated pressure displacement amount calculation unit 130. The processing shown in FIG. 3 is repeatedly executed at the same cycle as the processing in FIG. 2.

[0031] According to FIG. 3, first, the integrated pressure displacement amount calculation unit 130 acquires the operation data (operating / stopped) of the air compressor 210, the BC pressure value of the brake device 220, and the AS pressure value of the air spring 230 from the data input unit 110 (step S201). As described in the explanation of FIG. 2, the processing in step S201 is executed at the same cycle as the processing in step S101 of FIG. 2.

[0032] Next, the integrated pressure displacement calculation unit 130 refers to the operation data acquired in step S201 and determines whether the operating state of the air compressor 210 has changed from the stopped state to the operating state (step S202). If the operating state has changed from the stopped state to the operating state (YES in step S202), the process proceeds to step S204. If the operating state has not changed from the stopped state to the operating state (NO in step S202), the process proceeds to step S203.

[0033] In step S203, the integrated pressure displacement calculation unit 130 determines whether the operating state of the air compressor 210 is in the operating state based on the operation data acquired in step S201. If the operating state is in the operating state (YES in step S203), the process proceeds to step S204. If the operating state is not in the operating state (NO in step S203), the process proceeds to step S206 described later.

[0034] In step S204, the integrated pressure displacement calculation unit 130 calculates the integrated pressure displacement for the pressure data of each related device. Specifically, the integrated pressure displacement calculation unit 130 calculates the difference between the BC pressure value and the AS pressure value acquired in step S201 and the pressure value one cycle before for each, and integrates the absolute value of this difference into the integrated pressure displacement of each to calculate the integrated pressure displacement of the BC pressure value (integrated BC pressure displacement) and the integrated pressure displacement of the AS pressure value (integrated AS pressure displacement). Then, the integrated pressure displacement calculation unit 130 stores the current BC pressure value and AS pressure value in a predetermined storage area (step S205) and proceeds to step S206.

[0035] In step S206, the integrated pressure displacement calculation unit 130 determines whether the operating state of the air compressor 210 has changed from the operating state to the stopped state based on the operation data acquired in step S201. If the operating state has changed from the operating state to the stopped state (YES in step S206), the process proceeds to step S207. If the operating state has not changed from the operating state to the stopped state (NO in step S206), the process ends.

[0036] In step S207, the integrated pressure displacement calculation unit 130 outputs the current integrated BC pressure displacement and the integrated AS pressure displacement to the operation time prediction model unit 140. Then, the integrated pressure displacement calculation unit 130 initializes the value of the integrated pressure displacement of each related device it manages to "0" (step S208) and ends the process.

[0037] As described above, through the calculation process of the integrated pressure displacement shown in FIG. 3, the integrated pressure displacement calculation unit 130 can calculate the amount of compressed air used by each related device for one operation from the start to the stop of the air compressor 210.

[0038] In this example, the case where the pressure data input from the data input unit 110 to the integrated pressure displacement calculation unit 130 is the BC pressure and the AS pressure has been described. However, when the pressure data input to the data input unit 110 is other than the above, the integrated pressure displacement calculation unit 130 may perform the processes of steps S204, S205, S207, and S208 for each input pressure data.

[0039] (1 - 4) Processing of the operation time prediction model unit 140 FIG. 4 is a flowchart showing an example of the processing procedure by the operation time prediction model unit 140. The processing shown in FIG. 4 starts when the integrated pressure displacement of each related device is output from the integrated pressure displacement calculation unit 130 in step S207 of FIG. 3.

[0040] According to FIG. 4, first, the operation time prediction model unit 140 acquires the integrated pressure displacement of each related device (for example, the integrated BC pressure displacement and the integrated AS pressure displacement) output from the integrated pressure displacement calculation unit 130 in step S207 of FIG. 3 (step S301).

[0041] Next, the operation time prediction model unit 140 substitutes the integrated pressure displacement acquired in step S301 into the prediction model formula of Formula 1 described later to calculate the predicted operation time (step S302).

[0042] Then, the operating time prediction model unit 140 outputs the predicted operating time calculated in step S302 to the fault sign diagnosis unit 150 (step S303), and ends the process.

[0043] (1-4-1) Prediction model A prediction model for calculating the predicted operating time of the air compressor 210 will be described in detail. The prediction model formula used in step S302 of FIG. 4 is shown by, for example, the following formula 1. [Number] In formula 1, Y t [seconds] is the predicted operating time, A [seconds] is the reference operating time, X n [kPa] is the integrated pressure displacement amount of the related device n, and α n [seconds / kPa] is the influence degree of the related device n.

[0044] The reference operating time A is the normal operating time of the air compressor 210 in a configuration not affected by related devices. The reference operating time A is the normal air compression performance W [kPa / second] of the air compressor 210 and the starting pressure P min and the stop pressure P max are known in advance, then it can be calculated using the formula "A=(P max -P min ) / W". Note that the starting pressure P min is the lower pressure value of the starting pressure in the above tank, and the stop pressure P max is the upper pressure value of the stop pressure in the above tank.

[0045] Also, when the values of the air compression performance W, the starting pressure P min , and the stop pressure P max are unknown, the actual operating time values when the integrated pressure displacement amount of each related device is "0" during a certain period from when it is new or immediately after inspection or repair can be accumulated, and the average value thereof can be used. The above certain period is a period during which data that comprehensively includes the operations (between stations, types) assumed in the vehicle (train) on which the air compressor 210 is mounted can be acquired.

[0046] Degree of influence α of each related device n is a parameter indicating the degree of influence on the air compressor 210 of the integrated pressure displacement amount in each related device. More specifically, it is a parameter indicating the extension amount [seconds] of the operation time of the air compressor 210 per 1 kPa of compressed air used by each related device. The degree of influence α n for each value α 1 ~α n is calculated in advance using regression learning or the like with the actual operation time for a certain period and the data of the integrated BC pressure displacement amount and the integrated AS pressure displacement amount at that time as learning data.

[0047] Also, when the integrated pressure displacement amount other than the related device n is "0", as shown in FIG. 5, based on the relationship between the integrated pressure displacement amount of the related device n and the operation time at that time, the degree of influence α n of the related device n may be determined. Specifically, for example, the integrated pressure displacement amount of the related device n and the corresponding operation time are accumulated over a certain period, an approximate formula is calculated using a known method such as the least squares method, and the value of the slope in the approximate formula is determined as the degree of influence α n (see FIG. 5). The above-mentioned certain period is a period during which data comprehensively including the operation (between stations, type) assumed in the vehicle (train) equipped with the air compressor 210 can be acquired.

[0048] FIG. 5 is a diagram for explaining an example of a method for determining the degree of influence α n of a related device. As shown in FIG. 5, an approximate formula can be calculated for the data of the operation time and the integrated pressure displacement amount of the related device n accumulated over a certain period using a predetermined method. Then, the slope in the approximate formula of the linear function shown in FIG. 5 can be determined as the degree of influence α n of the related device n.

[0049] (1-5) Processing of the fault prediction diagnosis unit 150 FIG. 6 is a flowchart showing an example of a processing procedure of the processing by the fault prediction diagnosis unit 150. The processing shown in FIG. 6 is executed after the processing of FIGS. 2 and 4 is completed.

[0050] According to FIG. 6, first, the failure omen diagnosis unit 150 acquires the actual operation time from the operation time calculation unit 120 and acquires the predicted operation time from the operation time prediction model unit 140 (step S401).

[0051] Next, based on the actual operation time and the predicted operation time acquired in step S401, the failure omen diagnosis unit 150 determines whether the actual operation time is greater than the value obtained by adding the threshold d 1 to the predicted operation time (step S402). Although details will be described later, the threshold d 1 is a parameter used to determine whether a failure may occur in the air compressor 210 or its related equipment based on the actual operation time and the predicted operation time of the air compressor 210.

[0052] If an affirmative result is obtained in step S402 (YES in step S402), since the actual operation time of the air compressor 210 is too long compared to the predicted operation time, it can be determined that an abnormal (failure) state may occur in the air compressor 210 or related equipment. Therefore, the failure omen diagnosis unit 150 increments the value of the failure omen counter, which is a parameter it holds, by 1 (step S403) and proceeds to step S404.

[0053] On the other hand, if a negative result is obtained in step S402 (NO in step S402), since the actual operation time of the air compressor 210 is within the normal range compared to the predicted operation time, no omen of an abnormal (failure) occurring in the air compressor 210 or related equipment is detected. Therefore, the failure omen diagnosis unit 150 initializes the value of the failure omen counter to "0" (step S406) and proceeds to step S407.

[0054] In step S404, the failure omen diagnosis unit 150 determines whether the value of the failure omen counter is greater than the threshold d 2 (step S404). Although details will be described later, the threshold d 2is a parameter used to determine whether there is a sign of failure in the air compressor 210 or its related equipment from the value of the failure prediction counter.

[0055] In step S404, when the value of the failure prediction counter is greater than the threshold value d 2 (YES in step S404), it means that the value of the failure prediction counter has reached the detection standard of the failure prediction. Therefore, the failure prediction diagnosis unit 150 notifies the fact that the failure prediction has been detected to a predetermined notification means or its control means, and the predetermined notification means or control means notifies a warning (failure prediction alert) corresponding to the detection of the failure prediction to a predetermined notification destination such as an operator, an operation command, or an inspection area (step S405). The details of the notification of the failure prediction alert will be described later. After step S405, the process proceeds to step S407.

[0056] On the other hand, in step S404, when the value of the failure prediction counter is less than or equal to the threshold value d 2 (NO in step S404), it means that although the value of the failure prediction counter is 1 or more, it has not reached the detection standard of the failure prediction. Therefore, the failure prediction diagnosis unit 150 does not perform the notification of the failure prediction alert at the current stage and proceeds to step S407.

[0057] Then, in step S407, the failure prediction diagnosis unit 150 saves the current value of the failure prediction counter in a predetermined storage area, and then ends the process.

[0058] (1-5-1) Threshold value d 1 Threshold value d 1 The details of will be described. Threshold value d 1 is a parameter for determining whether the actual operation time is an abnormal value with respect to the predicted operation time, and indicates the standard of the excess time from the predicted operation time allowed for the actual operation time.

[0059] Threshold value d 1As a calculation method, for example, when calculating the reference operation time A in the above formula 1, if the average value of the accumulated data is used, the standard deviation σ of the accumulated data is calculated, and a value regarded as an outlier in statistics (such as 3σ) is used as the threshold d 1 may be used.

[0060] Also, for example, regarding the normal air compression performance W [kPa / second] of the air compressor 210 in advance, the maximum value W max and the minimum value W min are known, the threshold d 1 may be calculated from the following formula 2.

Number

[0061] Also, for the starting pressure P min and the stopping pressure P max of the air compressor 210, if it is known that there is a width, for example, like "P min ±ΔP 1 " and "P max ±ΔP 2 ", when the values of ΔP 1 , ΔP 2 are used, the threshold d 1 may be calculated from the following formula 3.

Number

[0062] Also, when the maximum value W max and the minimum value W min of the air compression performance W, and the above ΔP 1 , ΔP 2 are all known, the threshold d 1 may be calculated from the following formula 4.

Number

[0063] (1-5-2) Threshold d 2 Threshold d2 will be described in detail. Threshold d 2 is a parameter for determining whether a phenomenon in which the actual operation time is excessively long with respect to the predicted operation time (specifically, a phenomenon in which the actual operation time is longer than the threshold d 1 or more) occurs continuously, and indicates the standard for the number of consecutive occurrences of the above phenomenon.

[0064] The value of threshold d2 is determined from the train operation plan so that when the actual operation time continuously exceeds the predicted operation time by more than threshold d1 for a long time in data comprehensively including conditions such as between stations and time zones, it can be detected that a fault omen has occurred.

[0065] Also, in order to detect sudden failures such as when an object collides with an air pipe and a crack occurs, when the actual operation time greatly exceeds the value obtained by integrating the predicted operation time with threshold d 1 (for example, a value twice the threshold d 1 etc.), the value of threshold d 2 may be set to a smaller number of times (for example, 2 times, 3 times, etc.) than the value of the above-mentioned threshold d 2 .

[0066] (1-5-3) Notification of Fault Omen Alert The details of the method for notifying a fault omen alert will be described. When the fault omen diagnosis unit 150 detects a fault omen of the air compressor 210, the fault omen alert is a function in which a predetermined notification means and its control means notify the driver, the operation command, and the maintenance area of the detection of the fault omen of the air compressor 210 by means of screen display or sound emission.

[0067] For example, when the fault prediction diagnosis device 100 is installed on a train, a fault prediction alert is output to a vehicle information control device (not shown), and the vehicle information control device displays a screen indicating that a fault prediction has been detected on the display in the driver's cab. Also, the fault prediction diagnosis device 100 may directly display a screen indicating that a fault prediction has been detected on the display in the driver's cab. Further, when the fault prediction diagnosis device 100 has a function of transmitting aggregated vehicle data of the vehicle information control device to a ground system such as a driving command or an inspection area using public wireless communication or the like, the fault prediction alert may be included in the vehicle data and transmitted to the ground system, so that the fault prediction alert is notified through a screen or the like in the ground system. Furthermore, when the fault prediction diagnosis device 100 is installed on a ground system such as a driving command or an inspection area, the fault prediction alert may be displayed on a screen or the like on the ground system.

[0068] As described above, according to the fault prediction diagnosis device 100 according to the first embodiment, even when the air compressor 210 operates during the operation of the vehicle (train), it is possible to accurately estimate the operation time (predicted operation time) of the air compressor 210. Further, the fault prediction diagnosis device 100 has a function of calculating the amount of compressed air used by related devices that use the compressed air generated by the air compressor 210 during the operation of the air compressor 210, and a function of considering the degree of influence of the amount of compressed air used by each related device on the predicted operation time. By having these functions, the fault prediction diagnosis device 100, over a certain period (up to the number of times of the threshold d 2 continuously), when the actual operation time of the air compressor 210 shows a value longer than the predicted operation time by more than the threshold d 1 it can be determined that a fault prediction has occurred in the air compressor 210 or its related devices (for example, the brake device 220 or the air spring 230).

[0069] Thus, the fault prediction diagnostic device 100 according to the present embodiment can always accurately detect and diagnose the signs of abnormalities (faults) in the air compressor 210 that operates at various timings during the running of the vehicle, or in each related device that uses the compressed air generated by the air compressor 210, without being restricted by the operation and line conditions of the vehicle.

[0070] (2) Second Embodiment (2-1) Configuration FIG. 7 is a block diagram showing a configuration example of the fault prediction diagnostic device 100 according to the second embodiment of the present invention. As shown in FIG. 7, the configuration of the fault prediction diagnostic device 100 according to the second embodiment is the same as the configuration shown in FIG. 1 in the first embodiment. As a difference from the first embodiment, in the second embodiment, it is assumed that the vehicle on which the air compressor 210 is mounted has a pneumatic door device 240 that performs the opening and closing operation of the door using compressed air. That is, in the second embodiment, as related devices that use the compressed air generated by the air compressor 210, the door device 240 is added to the brake device 220 and the air spring 230 similar to those in the first embodiment, and data is input from the door device 240 to the data input unit 110.

[0071] Here, in general railway vehicles, the pressure value of the air tank in which the compressed air used by the door device 240 is accumulated is often not measured. In this case, pressure data cannot be obtained from the door device 240. Therefore, in the second embodiment, the data input unit 110 acquires the opening and closing information of the door as the input data from the door device 240.

[0072] In the fault prediction diagnostic device 100 according to the second embodiment configured as described above, since the operation time calculation unit 120 and the fault prediction diagnostic unit 150 are the same as those in the first embodiment in terms of both configuration and processing, the description thereof is omitted. Also, the data input unit 110, the integrated pressure displacement amount calculation unit 130, and the operation time prediction model unit 140 will be described centering on the differences from the first embodiment.

[0073] (2-2) Processing of the Data Input Unit 110 In the second embodiment, the data input unit 110 inputs operation data from the air compressor 210, pressure data (BC pressure value, AS pressure value) from the brake device 220 and the air spring 230, and opening / closing data from the door device 240. The opening / closing data is information indicating the opening / closing state of the door in the door device 240. Further, the data input unit 110 outputs the operation data of the air compressor 210 to the operation time calculation unit 120 and the integrated pressure displacement calculation unit 130, and outputs the pressure data (BC pressure value, AS pressure value) of the brake device 220 and the air spring 230, and the opening / closing data (open / close) of the door device 240 to the integrated pressure displacement calculation unit 130.

[0074] (2-3) Processing of the integrated pressure displacement calculation unit 130 The processing of the integrated pressure displacement calculation unit 130 in the second embodiment will be described with reference to FIG. 8. FIG. 8 is a flowchart showing an example of the processing procedure of the processing by the integrated pressure displacement calculation unit 130 of the second embodiment. In FIG. 8, for the same processing as that in FIG. 3 described in the first embodiment, the same reference numerals are given and the detailed description thereof is omitted.

[0075] According to FIG. 8, first, the integrated pressure displacement calculation unit 130 acquires, from the data input unit 110, the operation data (operating / stopped) of the air compressor 210, the BC pressure value of the brake device 220, the AS pressure value of the air spring 230, and the opening / closing data of the door device 240 (step S501).

[0076] The processing of the next steps S202 to S205 is the same as the processing of steps S202 to S205 in FIG. 3. After the processing of step S205, the process proceeds to step S502.

[0077] In step S502, the integrated pressure displacement calculation unit 130 determines whether or not the opening / closing state of the door device 240 has changed from closed to open based on the opening / closing data acquired in step S201. If the opening / closing state has changed from closed to open (YES in step S502), the process proceeds to step S503, and if the opening / closing state has not changed from closed to open (NO in step S502), the process proceeds to step S206.

[0078] In step S503, the integrated pressure displacement calculation unit 130 increments by 1 the value of the door opening count, which is a parameter it holds, and proceeds to step S206.

[0079] In step S206, as described with reference to FIG. 3, the integrated pressure displacement calculation unit 130 determines whether the operating state of the air compressor 210 has changed from the operating state to the stopped state based on the operation data acquired in step S201. If the operating state has changed from the operating state to the stopped state (YES in step S206), the process proceeds to step S504. If the operating state has not changed from the operating state to the stopped state (NO in step S206), the process ends.

[0080] In step S504, the integrated pressure displacement calculation unit 130 outputs the current integrated BC pressure displacement, the integrated AS pressure displacement, and the door opening count to the operation time prediction model unit 140. Thereafter, the integrated pressure displacement calculation unit 130 initializes the values of the integrated pressure displacement and the door opening count of each related device it manages to "0" (step S505), and ends the process.

[0081] (2-4) Processing of the operation time prediction model unit 140 In the second embodiment, the processing procedure of the operation time prediction model unit 140 is the same as the processing procedure example shown in FIG. 4 in the first embodiment. However, in the second embodiment, since the prediction model used is different, the points of change in the prediction model will be described below.

[0082] In the second embodiment, the operation time prediction model unit 140 can calculate the predicted operation time of the air compressor 210 using, for example, the following Equation 5.

Equation

[0083] In Equation 5, Y t [seconds] is the predicted operation time, A [seconds] is the reference operation time, X n [kPa] is the integrated pressure displacement of related device n, αn [sec / kPa] is the influence of the related device n. These are the same as the prediction model formula explained in the formula 1 in the first embodiment. Furthermore, in formula 5, the influence β 1 [Seconds / times] is the influence of the number of door openings, Z 1 [Times] is the number of times the door was opened.

[0084] Influence β 1 is a parameter indicating the extension amount [seconds] of the operation time of the air compressor 210 per one door opening in the door device 240. 1 is the influence degree α of the integrated pressure displacement amount described in the first embodiment. n It can be calculated in the same manner as in the above. Specifically, for example, it may be calculated in advance using regression learning or the like, in which the actual operation time for a certain period and the number of door openings at that time are used as learning data. Also, for example, as shown in FIG. 5, when the integrated pressure displacement amount of other related equipment is "0", the number of door openings in that situation and the operation time at that time are accumulated for a certain period, an approximation formula is calculated using a known method such as the least squares method, and the value of the slope in the approximation formula is used as the influence degree β 1 The above-mentioned certain period is a period during which data that comprehensively includes the expected operation (between stations, by type) of the vehicle (train) on which the air compressor 210 is mounted can be acquired.

[0085] As described above, according to the failure sign diagnosis device 100 of the second embodiment, the related devices included in the failure sign diagnosis target are not limited to only the first related devices from which pressure data can be obtained, such as the brake device 220 and the air spring 230, but even if a second related device from which pressure data cannot be obtained, such as the door device 240, is included, the predicted operation time of the air compressor 210 can be calculated taking into consideration the degree of influence of the first and second related devices on the operation time of the air compressor 210. That is, according to the failure sign diagnosis device 100 of the second embodiment, in a vehicle equipped with various related devices, it is possible to accurately estimate the operation time and always accurately diagnose the failure sign of the air compressor 210 or its related devices based on the estimated operation time without being limited by the operation and linear conditions of the vehicle.

[0086] In the above description, only the door device 240 is shown as the second related device that cannot acquire pressure data. However, even when a plurality of second related devices are connected, the failure prediction diagnostic apparatus 100 according to the second embodiment can similarly diagnose a failure prediction with high accuracy. In such a case, specifically, the predicted operation time may be calculated using Expression 5 in which the number of terms of "β × Z" corresponding to each second related device is added, and the failure prediction may be detected and diagnosed based on the calculated predicted operation time and the actual operation time.

[0087] (3) Third Embodiment A third embodiment of the present invention will be described. In the abnormality detection method disclosed in Patent Document 1 described above, a boundary line serving as a criterion for abnormality detection determination was calculated using data for one year from the start of business of a specific vehicle. However, in such a method, individual differences between vehicles and differences in characteristics when traveling on different sections even for vehicles of the same type cannot be considered, and there is a risk that the diagnostic accuracy may decrease when applied to vehicles other than the vehicle for which the boundary line was calculated. In view of such problems, in the third embodiment of the present invention, a failure prediction diagnostic apparatus is provided that can accurately detect and diagnose a failure prediction of an air compressor or its related devices without being affected by individual differences between vehicles and differences in the sections where the vehicles are introduced.

[0088] (3-1) Configuration FIG. 9 is a block diagram showing a configuration example of a failure prediction diagnostic apparatus 300 according to the third embodiment of the present invention. As shown in FIG. 9, the failure prediction diagnostic apparatus 300 includes a data input unit 310, an operation time calculation unit 320, an integrated pressure displacement amount calculation unit 330, an operation time prediction model unit 340, a failure prediction diagnosis unit 350, and a model update unit 360.

[0089] Further, in the fault prediction diagnosis device 300 of FIG. 9, similar to the fault prediction diagnosis device 100 shown in FIG. 1 in the first embodiment, an air compressor 210, a brake device 220, and an air spring 230 are connected as devices to be diagnosed for fault prediction. Note that FIG. 9 is an example, and the related devices of the air compressor 210 connected to the fault prediction diagnosis device 300 may be one or a plurality. Also, the related devices may be devices that cannot acquire pressure data, such as the door device 240 shown in the second embodiment.

[0090] The data input unit 310 has the same configuration as the data input unit 110 described in the first embodiment.

[0091] The operating time calculation unit 320 has the same configuration as the operating time calculation unit 120 described in the first embodiment, except that the calculated actual operating time is output to the fault prediction diagnosis unit 350 and the model update unit 360.

[0092] The integrated pressure displacement calculation unit 330 has the same configuration as the integrated pressure displacement calculation unit 130 described in the first embodiment.

[0093] The operating time prediction model unit 340 inputs the integrated pressure displacements (integrated BC pressure displacement, integrated AS pressure displacement) of each related device from the integrated pressure displacement calculation unit 330, and inputs the influence degree values of each integrated pressure displacement from the model update unit 360. Based on the input information, the predicted operating time of the air compressor 210 is calculated using a predetermined prediction model. Further, the operating time prediction model unit 340 outputs the calculated predicted operating time to the fault prediction diagnosis unit 350, and outputs the calculated predicted operating time, the integrated pressure displacements of each related device in the prediction model used for calculating the predicted operating time, and the influence degree values, to the model update unit 360. Details of the processing by the operating time prediction model unit 340 will be described later.

[0094] The fault prediction diagnosis unit 350 receives the actual operation time from the operation time calculation unit 320 and the predicted operation time from the operation time prediction model unit 340, and performs fault prediction diagnosis based on the input information. Further, the fault prediction diagnosis unit 350 outputs the diagnosis result of the fault prediction to the driver or the vehicle base, and outputs the value of the diagnosis state to the model update unit 360. In this description, the diagnosis state takes one of the values "0", "1", and "2" according to the state, but other values indicating other states may be prepared. The details of the processing by the fault prediction diagnosis unit 350 will be described later.

[0095] The model update unit 360 receives the actual operation time from the operation time calculation unit 320, the values of the influence degrees and feature quantities in the prediction model and the predicted operation time from the operation time prediction model unit 340, and the value of the diagnosis state from the fault prediction diagnosis unit 350. Then, based on the input information, the model update unit 360 updates the values of the influence degrees in the prediction model, and outputs the updated values of the influence degrees to the operation time prediction model unit 340. The details of the processing by the model update unit 360 will be described later.

[0096] In the fault prediction diagnosis device 300 configured as described above, the processing by the data input unit 310, the operation time calculation unit 320, and the integrated pressure displacement amount calculation unit 330 is the same as that in the first embodiment, so detailed description thereof is omitted. Hereinafter, the operation time prediction model unit 340, the fault prediction diagnosis unit 350, and the model update unit 360 that perform processing different from that in the first embodiment will be described in detail.

[0097] (3-2) Processing of the operation time prediction model unit 340 FIG. 10 is a flowchart showing an example of the processing procedure of the operation time prediction model unit 340.

[0098] According to FIG. 10, first, the operation time prediction model unit 340 acquires the integrated pressure displacement amounts of the related devices (for example, the integrated BC pressure displacement amount and the integrated AS pressure displacement amount) output from the integrated pressure displacement amount calculation unit 330, and the values α' of the influence degrees corresponding to the integrated pressure displacement amounts from the model update unit 360 1 ~α' nTo obtain it (step S601).

[0099] Next, the operating time prediction model unit 340 updates the influence degrees α1 to αn in the prediction model formula of Equation 1 to the values α’ 1 ~α’ n obtained in step S601 (step S602).

[0100] Next, the operating time prediction model unit 340 substitutes the integrated pressure displacement amount obtained in step S601 into the prediction model formula of Equation 1 to calculate the predicted operating time (step S603).

[0101] Then, the operating time prediction model unit 340 outputs the predicted operating time calculated in step S603 to the fault omen diagnosis unit 150, and outputs the predicted operating time, the integrated pressure displacement amounts X1 to Xn of the respective related devices used in the calculation of the predicted operating time, and the influence degrees α1 to αn to the model update unit 360 (step S604), and ends the process.

[0102] (3-3) Processing of the fault omen diagnosis unit 350 FIG. 11 is a flowchart showing an example of a processing procedure of the processing by the fault omen diagnosis unit 350. In FIG. 11, for the same processing as that in FIG. 6 described in the first embodiment, the same reference numerals are given and the detailed description thereof is omitted.

[0103] According to FIG. 11, first, the fault omen diagnosis unit 350 acquires the actual operating time from the operating time calculation unit 320 and acquires the predicted operating time from the operating time prediction model unit 340 (step S401).

[0104] Next, the fault omen diagnosis unit 350 determines whether the actual operating time is greater than the value obtained by adding the threshold d 1 to the predicted operating time based on the actual operating time and the predicted operating time acquired in step S401 (step S402). The threshold d 1 is a parameter calculated in the same manner as the threshold d 1 described in the first embodiment. In step S402, when the actual operating time is greater than the predicted operating time plus the threshold d 1If the value is greater than the sum (YES in step S402), the fault prediction diagnosis unit 350 increments the value of the fault prediction counter by 1 (step S403) and proceeds to step S701.

[0105] In step S701, the fault prediction diagnosis unit 350 sets the value of the diagnosis state to "0". The diagnosis state of "0" indicates a state where a fault prediction has been detected in the current cycle and it is necessary to continuously monitor whether the fault prediction will continue to occur in the future for diagnosis.

[0106] Subsequent to step S701, the fault prediction diagnosis unit 350 determines whether the value of the fault prediction counter is greater than the threshold value d 2 (step S404). If the value of the fault prediction counter is greater than the threshold value d 2 in step S404 (YES in step S404), the fault prediction diagnosis unit 350 notifies a fault prediction alert to a predetermined notification destination such as an operator, a driving command, or a maintenance area (step S405) and proceeds to step S407. On the other hand, if the value of the fault prediction counter is less than or equal to the threshold value d 2 in step S404 (NO in step S404), the fault prediction alert is not notified and the process proceeds to step S407. The details of the determination in step S404 and the processing according to the determination result are as described in the first embodiment.

[0107] If the actual operation time is less than or equal to the sum of the predicted operation time and the threshold value d 1 in step S402 (NO in step S402), the process proceeds to step S702.

[0108] In step S702, the fault prediction diagnosis unit 350 determines whether the value of the fault prediction counter is greater than "0". If the value of the fault prediction counter is greater than "0", that is, if it is "1" or "2" (YES in step S702), the fault prediction diagnosis unit 350 sets the value of the diagnosis state to "2" (step S703) and proceeds to step S406. The diagnosis state of "2" indicates a state where there may be a false detection because the fault prediction detected in the previous cycle was not detected in this cycle. On the other hand, if the value of the fault prediction counter is "0" (NO in step S702), the fault prediction diagnosis unit 350 sets the value of the diagnosis state to "1" (step S704) and proceeds to step S406. The diagnosis state of "1" indicates a normal state where no fault prediction has been detected following the previous cycle, or a state of false detection where the fault prediction has not been detected (also referred to as "undetected" to distinguish it from other false detections).

[0109] In step S406, the fault prediction diagnosis unit 350 initializes the value of the fault prediction counter to "0" and proceeds to step S407.

[0110] In step S407, the fault prediction diagnosis unit 350 stores the current value of the fault prediction counter in a predetermined storage area.

[0111] Finally, the fault prediction diagnosis unit 350 outputs the current value of the diagnosis state to the model update unit 360 (step S705) and ends the process.

[0112] (3-4) Processing of the model update unit 360 FIG. 12 is a flowchart showing an example of the processing procedure of the processing by the model update unit 360. The processing in FIG. 12 starts after the processing in step S705 of FIG. 11.

[0113] According to FIG. 12, first, the model update unit 360 acquires the actual operation time from the operation time calculation unit 320, acquires the predicted operation time from the operation time prediction model unit 340, the integrated pressure displacement amount and the influence degree of each related device, and acquires the diagnosis state from the fault omen diagnosis unit 350 (step S801).

[0114] Next, the model update unit 360 determines the value of the diagnosis state acquired in step S801 (step S802). The value of the diagnosis state is either "0", "1", or "2". If the value of the diagnosis state is "0", the process proceeds to step S803. If the value of the diagnosis state is "1", the process proceeds to step S804. If the value of the diagnosis state is "2", the process proceeds to step S807.

[0115] When the diagnosis state is "0", a fault omen is detected in the current cycle, and it is in a state of continuously monitoring whether it will continue to occur in the future. In this state, there is no need to change the parameters of the prediction model that detects the fault omen. Therefore, in step S803, the model update unit 360 updates the values of the predicted operation time, the actual operation time, and the integrated pressure displacement amount of each related device stored in its memory area to the values acquired in step S801, respectively, and ends the process.

[0116] When the diagnosis state is "1", it is a normal state in which no fault omen is detected, or a state of false detection (undetected) in which no fault omen can be detected. This state of false detection (undetected) means that due to individual differences in the vehicle, etc., a predicted operation time longer than expected is calculated, and thus no fault omen can be detected (undetected). Considering these states, in step S804, the model update unit 360 determines whether the predicted operation time is greater than the value obtained by adding the threshold d 1 to the actual operation time. When the predicted operation time is greater than the value obtained by adding the threshold d 1 to the actual operation time (YES in step S804), since it represents a state where the predicted operation time is too long and there is a risk of undetected, the process proceeds to step S805. On the other hand, when the predicted operation time is less than or equal to the value obtained by adding the threshold d 1If the value added is less than or equal to (NO in step S804), there is no problem with the predicted operation time and no changes are necessary, so the process ends.

[0117] In step S805, the model update unit 360 calculates a value X obtained by multiplying the integrated pressure displacement amount of each related device acquired in step S801 by the corresponding influence degree. 1 α 1 ~X n α n and calculates the ratio γ 1 ~γ n of the magnitudes of the calculated values using the following formula 6 and proceeds to step S806.

Equation

[0118] In step S806, the model update unit 360 updates the value of the influence degree α 1 ~γ n according to the ratio γ 1 ~α n of each related device calculated in step S805 to the influence degree α' 1 ~α' k using the following formula 7 and proceeds to step S809.

Equation

[0119] When the diagnosis status is "2", it is a state where there is a possibility that the fault omen detected in the previous cycle was a false detection. In this state, it is preferable to change the prediction model for detecting the fault omen, and it is necessary to calculate the updated value of the influence degree of the prediction model. Therefore, in step S807, the model update unit 360 multiplies the integrated pressure displacement amount of each related device stored in its memory area by the corresponding influence degree to obtain a value X 1 α 1 ~X n α n and calculates the ratio γ 1 ~γ n of the magnitudes of the calculated values, and proceeds to step S808 using the above-mentioned formula 6.

[0120] In step S808, the model update unit 360 updates the value of the influence degree α 1 ~γ n to the influence degree α' 1 ~α n according to the ratio γ 1 ~α’ k of each related device calculated in step S807, and proceeds to step S809.

Equation

[0121] Then, in step S809, the model update unit 360 outputs the updated influence degree α' k of related device k calculated in step S806 or step S808 to the operation time prediction model unit 340, and ends the process.

[0122] As described above, in the fault prediction diagnostic device 300 according to the third embodiment, the diagnostic state is determined based on the magnitude relationship between the actual operation time and the predicted operation time and the detection status of the fault prediction, and the parameters of the prediction model (the degree of influence of each related device on the operation time of the air compressor 210) are sequentially updated, so that it is possible to update to a prediction model that takes into account the individual differences between vehicles and the differences in the line sections where the vehicles are introduced. Thus, the fault prediction diagnostic device 300 according to the third embodiment can accurately estimate the operation time of the air compressor 210 regardless of the individual differences between vehicles and the differences in the line sections where the vehicles are introduced, and using this estimated operation time, it is possible to accurately detect and diagnose the fault prediction of the air compressor 210 or its related devices.

[0123] In the above description of the third embodiment, similar to the first embodiment, the brake device 220 and the air spring 230 capable of acquiring pressure data are used as related devices. However, even when related devices that cannot acquire pressure data are connected as in the second embodiment, it is possible to implement the third embodiment.

[0124] Note that the present invention is not limited to the above-described embodiments and includes various modifications. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described. Other aspects conceivable within the scope of the technical idea of the present invention are also included in the scope of the present invention. Also, it is possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. Further, it is possible to add, delete, or replace a part of the configuration of each embodiment with another configuration.

[0125] Further, each of the above configurations, functions, processing units, processing means, etc. may be realized in hardware by designing a part or all of them, for example, by means of an integrated circuit. Also, each of the above configurations, functions, etc. may be realized in software by a processor interpreting and executing a program for realizing each function. Information such as programs, tables, files, etc. for realizing each function can be placed in a memory, a recording device such as a hard disk, SSD (Solid State Drive), or a recording medium such as an IC card, SD card, DVD, or may be in a form in which a program etc. is distributed from an external server or cloud.

[0126] Also, in the drawings, control lines and information lines show those considered necessary for explanation, and not necessarily all control lines and information lines are shown on the product. In reality, it may be considered that almost all components are interconnected.

Explanation of Signs

[0127] 100, 300 Fault prediction diagnosis device 110, 310 Data input unit 120, 320 Operating time calculation unit 130, 330 Integrated pressure displacement calculation unit 140, 340 Operating time prediction model unit 150, 350 Fault prediction diagnosis unit 210 Pneumatic compressor 220 Brake device 230 Air spring 240 Door device 360 Model update unit

Claims

1. A fault prediction diagnosis device for diagnosing a fault prediction in a pneumatic compressor or related equipment using compressed air generated by the pneumatic compressor, a data input unit for inputting the operation data of the pneumatic compressor and the pressure data of the related equipment, an operation time calculation unit for calculating the actual operation time of the pneumatic compressor based on the operation data input to the data input unit, an integrated pressure displacement amount calculation unit for calculating, for each related equipment, an integrated pressure displacement amount indicating the amount of compressed air used by the related equipment for the operation of the pneumatic compressor based on the operation data and the pressure data input to the data input unit, an operation time prediction model unit for calculating a predicted operation time of the pneumatic compressor from a prediction model that uses, as a feature amount, the integrated pressure displacement amount of each related equipment calculated by the integrated pressure displacement amount calculation unit, a fault prediction diagnosis unit for diagnosing the presence or absence of a fault prediction in at least one of the pneumatic compressor or the related equipment based on a comparison result between the actual operation time calculated by the operation time calculation unit and the predicted operation time calculated by the operation time prediction model unit, A fault prediction diagnosis device characterized by comprising the above.

2. The fault prediction diagnosis unit, compares the actual operation time and the predicted operation time periodically, when, in the comparison, the actual operation time is longer than the total time of the predicted operation time and a first threshold value, detects that there is a possibility of the fault prediction, diagnoses the presence or absence of the fault prediction based on the detection result of the possibility of the fault prediction The fault prediction diagnosis device according to claim 1, characterized by the above.

3. The fault prediction diagnosis unit, when the detection of the possibility of the fault prediction continues for a number of times exceeding a second threshold value, diagnoses that there is a fault prediction The fault prediction diagnosis device according to claim 2, characterized by the above.

4. The prediction model calculates the predicted operation time by adding an additional operation time, which is the operation time of the pneumatic compressor added by the operation of the related equipment, to the normal operation time of the pneumatic compressor in a configuration not affected by the related equipment, The additional operation time is calculated by multiplying the feature amount of each related equipment by an influence degree determined by a predetermined method for each related equipment. The fault prediction diagnosis device according to claim 2 or claim 3, characterized by the above.

5. When the related devices include a first related device capable of acquiring pressure data and a second related device that cannot acquire pressure data but can acquire predetermined operation data, the operation time prediction model unit uses the integrated pressure displacement amount calculated based on the pressure data as a feature amount of the first related device, and the number of operations calculated based on the operation data as a feature amount of the second related device, and calculates the predicted operation time of the air compressor from the prediction model using the feature amounts of the first and second related devices. The failure omen diagnosis device according to any one of claims 1 to 4, characterized in that.

6. When one of the first related devices is a brake device, the pressure data of the first related device is a brake cylinder pressure value, When one of the first related devices is an air spring, the pressure data of the first related device is an air suspension pressure value, When one of the second related devices is a door device, the operation data of the second related device is door opening and closing information The failure omen diagnosis device according to claim 5, characterized in that.

7. The failure omen diagnosis device further includes a model update unit that updates the influence degree of each of the related devices in the prediction model based on the comparison result between the actual operation time and the predicted operation time. The failure omen diagnosis device according to claim 4, characterized in that.

8. The model update unit, When the actual operation time is less than or equal to the total time of the predicted operation time and the first threshold value, and the predicted operation time is longer than the total time of the actual operation time and the first threshold value, Reduce the influence degree of each of the related devices according to the proportion of each feature amount in the predicted operation time. The failure omen diagnosis device according to claim 7, characterized in that.

9. The failure omen diagnosis unit, Periodically compares the actual operation time and the predicted operation time, In the comparison, when the actual operation time is longer than the total time of the predicted operation time and the first threshold value, it is detected that there is a possibility of a failure omen, When the detection of the possibility of a failure omen continues a number of times exceeding a second threshold value, it is diagnosed that there is a failure omen, The prediction model calculates the predicted operation time by adding an additional operation time, which is the operation time of the air compressor added by the operation of the related device, to the normal operation time of the air compressor in a configuration not affected by the related device. The additional operating time is calculated for each of the related devices by multiplying the influence degree determined by a predetermined method on the feature amount of the related device, further comprising a model update unit that updates the influence degree of each of the related devices in the prediction model based on a comparison result between the actual operating time and the predicted operating time, The model update unit is when the actual operating time is equal to or less than the total time of the predicted operating time and the first threshold value, and the predicted operating time is longer than the total time of the actual operating time and the first threshold value, reducing the influence degree of each of the related devices according to the proportion of each feature amount at the predicted operating time, before the state in which the possibility of the fault sign is detected by the fault sign diagnosis unit continuously occurs more than the number of times of the second threshold value, when the actual operating time becomes equal to or less than the total time of the predicted operating time and the first threshold value, increasing the influence degree of each of the related devices according to the proportion of each feature amount at the predicted operating time The fault sign diagnosis device according to claim 1, characterized in that.

10. When the fault sign diagnosis unit diagnoses that there is a fault sign, it notifies an alert to a predetermined output destination The fault sign diagnosis device according to any one of claims 1 to 9, characterized in that.

11. A fault sign diagnosis method by a fault sign diagnosis device for diagnosing a fault sign in a pneumatic compressor or a related device that uses compressed air generated by the pneumatic compressor, a data input step in which the fault sign diagnosis device inputs the operation data of the pneumatic compressor and the pressure data of the related device, an operation time calculation step in which the fault sign diagnosis device calculates the actual operation time of the pneumatic compressor based on the operation data input in the data input step, an integrated pressure displacement amount calculation step in which the fault sign diagnosis device calculates, for each of the related devices, an integrated pressure displacement amount indicating the amount of compressed air used by the related device for the operation of the pneumatic compressor based on the operation data and the pressure data input in the data input step, an operation time prediction step in which the fault sign diagnosis device calculates the predicted operation time of the pneumatic compressor from a prediction model that uses the integrated pressure displacement amount of each related device calculated in the integrated pressure displacement amount calculation step as a feature amount A failure omen diagnosis step in which the failure omen diagnosis device diagnoses the presence or absence of a failure omen in at least one of the air compressor or the related equipment based on a comparison result between the actual operation time calculated in the operation time calculation step and the predicted operation time calculated in the operation time prediction step; A failure omen diagnosis method characterized by comprising the above.

Images