JP2006315813A - Movable body diagnosis system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To automatically diagnose a movable body, and prevent failure. <P>SOLUTION: A vibration sensor is installed on an overhead vehicle 16 to make input to a diagnosis device 26 in addition to output torque and rotation speed or the like of a travel motor and a lift motor. The diagnosis device 26 statistically processes obtained data to be analyzed by a support vector machine group 68 and a cluster analysis part 70. The overhead vehicle 16 possibly having abnormality is run in a maintenance area to determine if it is abnormality on the side of the overhead vehicle 16 or abnormality on the travel rail side. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to diagnosis of a moving body such as a tracked carriage such as an overhead traveling vehicle or a stacker crane, an article transfer device, a tool feeding device of a machine tool, and the like.

  The overhead traveling vehicle system will be described as an example. Since this system is mainly provided in a clean room, it has the property of disabling manual inspection. Since this system is installed near the ceiling, inspection is not easy. On the other hand, since 100 or more overhead traveling vehicles are operated in a large-scale system, in order to maintain the system, it is necessary to detect an abnormality early in order to prevent a failure of the overhead traveling vehicle.

  The overhead traveling vehicle is equipped with various encoders, sensors that detect the presence or absence of loads on the station on the ground, obstacle sensors, controllers and communication devices with other overhead traveling vehicles, etc. Detection is easy because it tends to appear as a control abnormality. In contrast, abnormalities in mechanisms such as motors, wheels, bearings, and cams, especially abnormalities that gradually accumulate due to secular change, can often be determined only after the overhead traveling vehicle is lowered to the ground and overhauled. Therefore, it is effective for preventing a full-scale failure, especially when an abnormality of a mechanical mechanism portion can be detected at an early stage while the overhead traveling vehicle is traveling without being lowered to the ground.

An object of the present invention is to automatically diagnose a moving body and prevent a failure.
An additional problem in the invention of claim 2 is that it is possible to automatically obtain a symptom of abnormality in addition to normal / abnormality of the moving body.
An additional problem in the invention of claim 3 is to not miss an abnormality which is different from known abnormality data.
It is an additional object of the invention of claim 4 to automatically diagnose a tracked carriage that is generally difficult to inspect manually.
An additional problem in the invention of claim 5 is that the number of inspections in a clean room or the like is reduced, and an abnormality of a plurality of overhead traveling vehicles is detected at an early stage. The purpose is to be able to discriminate whether the side equipment is abnormal.

  The present invention relates to a diagnostic system for a moving body that moves according to a predetermined speed pattern consisting of acceleration, constant speed movement, and deceleration along a finite and predetermined movement path. Monitoring means for monitoring the state of the body moving motor, and the process from the output of the sensor and the output of the monitoring means to the start and stop of movement of the moving body, or the acceleration range, constant speed range, deceleration range A feature amount calculation means for obtaining a plurality of feature amounts, each of which indicates a feature for each section divided into a plurality of feature amounts as vectors in a multi-dimensional vector space, and the normality of the moving object from the position in the vector space / A discriminating means for discriminating an abnormality is provided.

  The feature amount may be a statistical amount obtained by statistically analyzing the sensor signal or the monitoring unit signal, or may be obtained by processing these signals by a filter bank, a wavelet transform, a short-time Fourier transform, or the like. Only features having high correlation with sensor normality / abnormality may be used. However, it is possible to use feature values without evaluating the presence or absence of correlation by processing with a discriminator so that low correlation does not contribute. . The diagnostic system includes, for example, a moving body and a diagnostic device, and the feature amount is calculated by, for example, the diagnostic device, but may be calculated by the moving body. In a small-scale diagnostic system for one stacker crane, one machine tool, or the like, the diagnostic device may be built in the stacker crane or machine tool.

  Preferably, a plurality of the discriminating means are provided corresponding to the type of abnormality of the moving body.

  Also preferably, means for storing a cluster consisting only of the vector obtained from a normal moving body, and whether or not a vector consisting of a plurality of feature values inputted from the moving body belongs to the cluster. Cluster analysis means for discriminating is further provided.

  Particularly preferably, the track is a tracked carriage that travels along the traveling rail and moves the article up and down, and the normality / abnormality is diagnosed for at least the travel and lift of the tracked carriage. The tracked truck is, for example, an overhead traveling vehicle, a grounded tracked cart, or a stacker crane.

  Most preferably, the mobile body is a plurality of overhead traveling vehicles traveling on a traveling rail provided in a ceiling space in a clean room, and the overhead traveling vehicle monitors the state of vibration detection sensors, traveling motors, and lifting motors. The monitoring device determines whether the overhead traveling vehicle is normal or abnormal based on a vector composed of the sensor and the characteristic amount of the signal of the monitoring means, and determines whether the overhead traveling vehicle is abnormal. On the other hand, re-diagnosis is performed from data in other sections of the traveling rail.

  In this invention, the mechanical state of the moving body is checked by means of the state monitoring means of the vibration sensor and the moving motor. Since the moving body moves along a finite and predetermined moving path and moves according to a predetermined speed pattern consisting of acceleration, constant speed running, and deceleration, the signal of the sensor and monitoring means should be a repetition of the signal of the same pattern is there. For this reason, it is possible to determine whether the signal is normal or abnormal by converting the signal of the sensor or the monitoring means into a feature value and examining the position in the multidimensional vector space. In this way, the deviation from the normal state can be discriminated with high sensitivity, so that it can be automatically detected before the abnormality of the moving body develops into a failure. The present invention can be used to diagnose abnormalities in electrical / electronic devices such as sensors and communication devices, but is suitable for diagnosing abnormalities in mechanical or mechatronic devices such as moving motors, bearings, wheels, cams and sprockets. ing.

  According to the second aspect of the invention, not only the presence / absence of an abnormality, but also an abnormality symptom, in other words, an abnormality type is simultaneously determined.

  When abnormal data having a type different from known abnormal data occurs, the determination means may miss the abnormality. For example, when a determination unit is provided for each type of abnormality and a type of abnormality that does not include a determination unit occurs, a large amount cannot be expected of the determination unit. The discriminating means obtains a hyperplane that is a threshold between known abnormal data and normal data, for example, and it is doubtful whether abnormal data far from the known abnormal data can be detected. In such a case, the invention of claim 3 detects the data that does not belong to the cluster consisting only of normal data, so that unknown abnormal data is rarely overlooked.

  A tracked carriage that travels along a traveling rail and raises and lowers articles is a central device in automated warehouses and transport systems, and is generally difficult to inspect manually and must maintain a high operating rate. is there. According to the invention of claim 4, the abnormality of the tracked carriage can be automatically and early detected.

  According to the invention of claim 5, since the vehicle travels in a high place in a clean room, the normal / abnormality of the overhead traveling vehicle, which is particularly difficult to inspect manually, is diagnosed. You can also determine whether something is abnormal. For this reason, even an overhead traveling vehicle system having a large number of overhead traveling vehicles can be detected before an abnormality of the overhead traveling vehicle develops into a full-scale failure.

  The best mode for carrying out the present invention will be described below.

  An example is shown in FIGS. In the figure, reference numeral 2 denotes an overhead traveling vehicle system, which is a track system with a track traveling on the ground, a stacker crane for transferring a liquid crystal substrate, etc. in a clean room, a transfer device, or a tool tool and a tool box. A moving body such as a tool exchanger for exchanging tools may be used. The overhead traveling vehicle system 2 is provided using a ceiling space in a clean room, 4 is an inter bay route, 6 is an intra bay route, 8 is a maintenance bay route, and is a special type of the intra bay route 6. The maintenance bay route 8 is provided with a lap route 10 so that the overhead traveling vehicle 16 can lap around. Reference numeral 12 denotes a branching junction, and reference numeral 14 denotes a curve. Reference numeral 16 denotes an overhead traveling vehicle that travels in the direction of the arrow in FIG. 1. For example, about 100 to several hundreds of overhead traveling vehicles 16 exist in the system 2.

  A station 18 is provided in a semiconductor processing apparatus or the like, and exchanges articles with the overhead traveling vehicle 16. A buffer 20 is used for temporary storage of articles. A maintenance area 22 is provided in the maintenance bay route 8 so that maintenance such as inspection, disassembly, and repair of the overhead traveling vehicle 16 can be performed. Reference numeral 24 denotes an overhead traveling vehicle controller which transmits a conveyance command to the overhead traveling vehicle 16 and a command such as traveling to the maintenance bay route 8 and receiving maintenance. Reference numeral 26 denotes a diagnostic device that receives the state of the overhead traveling vehicle 16 via, for example, the controller 24 and performs diagnosis.

  FIG. 2 shows the overhead traveling vehicle 16 and its traveling rail 30. The inter-bay route and the intra-bay route use the same traveling rail 30. The traveling rail 30 includes an upper rail 31 and a lower rail 32, and a litz wire 33 is disposed on the lower rail 32 to perform non-contact power feeding to the power receiving unit 40 of the overhead traveling vehicle 16. Reference numeral 34 denotes a traveling carriage, which includes, for example, a drive wheel 35, a driven wheel 36, and a guide wheel 37, and travels by a travel motor 38. Reference numeral 39 denotes a vibration sensor which is attached to the traveling carriage 34 using, for example, an acceleration sensor and measures vibrations in the x, y and z directions independently.

  The power receiving unit 40 receives power from the litz wire 33 by non-contact power feeding and simultaneously communicates with the overhead traveling vehicle controller and other overhead traveling vehicles via the litz wire 33. Reference numeral 41 denotes a frame, and reference numeral 42 denotes a lateral feed unit, which is used when the turntable 44 to the lifting platform 48 are laterally fed to the side of the traveling rail 30 to deliver articles to and from the buffer. Reference numeral 43 denotes a transverse feed motor. Reference numeral 44 denotes a turntable, which includes a rotation motor 45, and rotates the elevating drive unit 46 and the elevating table 48 within a horizontal plane to rotate the elevating table in a direction suitable for delivery to a buffer or a station. Reference numeral 46 denotes an elevating drive unit that elevates the elevating platform 48 by winding / unwinding an elevating motor 47 and a suspension material (not shown).

  Reference numeral 49 denotes a cover, which is provided, for example, at the front and rear of the overhead traveling vehicle 16, and 50 is an article, which is a semiconductor cassette here. In the embodiment, servo driving is performed for each of the motors 38, 43, 45, and 47. Therefore, the output torque is obtained from the driving current and the motor rotation speed is monitored by an encoder provided on the motor shaft or the like. . In the embodiment, the vibration sensor 39 is provided on the traveling carriage 34. However, the vibration sensor 39 may be provided on other parts. For example, when the carriage 34 is provided on the lifting platform 48, the vibration received by the article 50 during conveyance can be measured.

  The operation of an overhead traveling vehicle is a combination of acceleration, constant speed motion, and deceleration. In the case of the traveling speed pattern 51 of FIG. 3, 52 is an acceleration section, 53 is a constant speed section, 54 is a deceleration section, and these are defined as one section, respectively, in the x direction, y direction, and z direction of the traveling carriage in each section. Each of these vibrations, the output torque of the traveling motor and the rotational speed are monitored. If the constant speed section 53 is too long, for example, if it is 1 second or longer, it may be divided into a plurality of sections.

  In the ascending / descending speed pattern 55, 56 lowers the lifting platform in the descending section, and after the section 56 is finished, operates the chuck of the lifting platform to deliver the article. Thereafter, it is the rising section 57 that raises the lifting platform. Since the descending section 56 and the ascending section 57 are generally as short as 1 second or less, they are divided into one section without being divided. In the lateral feed speed pattern 58, 59 is an advance section, and the vertical drive section is advanced to the side of the traveling rail by the lateral feed section, and the lift drive section is returned in the return section 60. Between the advancing section 59 and the return section 60, the operation of the lifting platform and the delivery of goods are performed. Since the advance section 59 and the return section 60 are each a short section, each of them is defined as one section.

  4 shows the configuration of the diagnostic device 26. The overhead traveling vehicle 16 transmits the degree of vibration in each of the x, y, and z directions, the torque of the motor, and the rotational speed of the motor. In addition to this, the overhead traveling vehicle 16 transmits the current position and time, as well as the details of the operation such as during traveling, during the raising and lowering of the elevator platform, and during lateral feed. The input interface 62 of the diagnostic device 26 receives these data from an overhead traveling vehicle controller or the like, for example, stores raw data in the storage unit 64, and obtains statistics for these data in the statistics unit 66. The storage unit 64 stores the raw data separately for each overhead traveling vehicle so that it can be read later. When the data from the overhead traveling vehicle is divided into a plurality of data and transmitted, they are connected to form a set of data. The storage unit 64 may store the statistics obtained by the statistics unit 66 instead of storing raw data.

  The statistics unit 66 converts the data of the overhead traveling vehicle into statistics. As statistics, for example, the variance, absolute value of the difference between the maximum value and the minimum value, the variance of the moving average, etc., such as vibration, motor torque, or motor speed are obtained for each section in FIG. In addition, an average value, a third-order or fourth-order moment, and the like may be obtained. Instead of the statistics unit 66, filters such as a filter bank, a wavelet transform unit, and a short-time Fourier transform unit may be provided to extract feature amounts.

  The obtained statistics are input to a group of SVM group 68 and cluster analysis unit 70. In the SVM group 68, an SVM (support vector machine) is provided for each symptom to be diagnosed. For example, if one guide symptom is that the guide roller of the traveling carriage of the overhead traveling vehicle is loose, the normal / guide roller is somewhat loose. Two SVMs can be diagnosed in three stages: loose / guide roller is further loose. For example, in the first SVM, the guide roller is distinguished from the case where the guide roller is not loosened, and the guide roller is slightly loosened and further loosened. In the second SVM, the guide roller is normal or slightly loosened, and Distinguish from loose things. By providing an SVM for each symptom to be diagnosed, an abnormality can be detected and the symptom can also be detected. SVM is provided for each symptom. For example, looseness of the guide roller and rattling of the driving wheel and driven wheel are different symptoms. The cluster analysis unit 70 extracts a cluster consisting only of normal data, and diagnoses that it is abnormal when data located outside the cluster is input. In this case, symptoms cannot be diagnosed.

  FIG. 5 schematically shows the discrimination with one SVM. The input data from the overhead traveling vehicle is multidimensional including vibration in each direction of x, y, z, motor rotation speed, motor torque, and the like. It can be considered as a vector. When it is necessary to increase the accuracy of diagnosis, a linear section of a traveling rail and a branching / merging portion or a curved section may be distinguished, and data indicating these types may be added to add the dimension of the type of traveling position. In addition, these dimensions may be added in consideration of an influence such as whether an empty cargo or an article is being conveyed. The vector space for diagnosis is a multidimensional space, and the SVM generates a hyperplane (a space with a one-dimensional dimension lower than the vector space) that most reliably discriminates between normal data and abnormal data. A hyperplane is generally represented by a linear combination of data of each dimension, and an SVM is automatically generated from teacher data input with normal / abnormal labels. Normal data and abnormal data can be discriminated from this hyperplane as a boundary, and normality / abnormality can be diagnosed from which side of the hyperplane the input data is.

  FIG. 6 schematically shows cluster analysis. For example, a set of only normal data is obtained, and an area including them is generated as a cluster. It is preferable to confirm that abnormal data is not included in the cluster. If abnormal data is included, deform the cluster to remove the abnormal data. It is determined whether or not the input vector is included in a cluster of normal data.

  The reason why the cluster analysis unit 70 is provided in addition to the SVM group 68 is that it is difficult to collect all abnormal modes and all types of abnormal data in advance. In the SVM group 68, it is difficult to detect an unexpected abnormal mode or to distinguish data that is different from past abnormal data as abnormal. Therefore, the cluster analysis unit 70 detects that data that is separated from the normal data in some sense is generated.

  Returning to FIG. 4, the diagnosis results from the SVM group 68 and the cluster analysis unit 70 are sent to the GUI (graphical user interface) 72 and the maintenance management unit 74, and the operator inputs via the GUI 72 and the processing in the maintenance management unit 74. Thus, the overhead traveling vehicle 16 is instructed to travel to the maintenance bay route 8 for inspection or to transmit diagnostic data again from another section. This instruction is given, for example, via the overhead traveling vehicle controller 24.

  When any abnormality is detected with respect to the overhead traveling vehicle 16, it may be unknown whether the abnormality is due to the overhead traveling vehicle or the abnormality caused by the traveling rail or the like. Therefore, if the re-diagnosis is performed based on the data in another section, or if the re-diagnosis is performed by running the maintenance beirut 8, it is possible to determine whether the overhead traveling vehicle is abnormal or the traveling rail is abnormal. From the teacher data input 76, the data obtained by the statistics unit 66 that has been diagnosed and can be distinguished from normal or abnormal is input to the SVM group 68 and the cluster analysis unit 70. As a result, the hyperplane in the SVM group 68 and the cluster in the cluster analysis unit 70 are gradually learned and changed.

In the embodiment, the following effects can be obtained.
(1) The state of overhead vehicles can be automatically diagnosed. For this reason, it is possible to minimize the entry of people into and out of the clean room for maintenance of the overhead traveling vehicle, and it is possible to detect an abnormality before a failure occurs.
(2) By using the SVM group 68, it is possible to determine the symptom along with the detection of abnormality.
(3) The cluster analysis unit 70 can reduce oversight even in an unexpected abnormal mode.
(4) It can detect minor abnormalities such as loose guide rollers, and it can also detect abnormalities that are difficult to detect such as motors, bearings, and wheels for movement. Therefore, abnormalities due to aging of these devices can be detected.
(5) It is possible to identify whether there is an abnormality on the traveling rail side or an abnormality on the overhead traveling vehicle side.
(6) Since the operation of the overhead traveling vehicle in each section such as acceleration, constant speed traveling, deceleration, etc. is substantially the same, the data that can be obtained under normal conditions regardless of whether the overhead traveling vehicle is different or the position is different. It is almost the same. For this reason, it is not necessary to provide an SVM group or cluster analysis unit for each overhead traveling vehicle, and data of a large number of overhead traveling vehicles can be processed by the same SVM group or cluster analysis unit.
(7) Diagnosis is facilitated by using statistics and features rather than raw data.

In the embodiment, the overhead traveling vehicle system 2 is taken as an example. However, a tracked cart system for ground traveling, a stacker crane for conveying an article in a clean room, a transfer device for transferring an article between the carriage and the station, etc. The same can be applied to. The same applies to a tool exchanger for exchanging a tool between a tool box of a machine tool and a blade base. Further, instead of SVM, another analysis method such as discriminant analysis may be used.

Plan view showing the layout of an overhead traveling vehicle system Front view with partially cutout showing overhead traveling vehicle and traveling rail Schematic diagram showing the speed pattern of traveling, raising and lowering, and lateral feed of the overhead traveling vehicle, and the section for feature amount extraction Block diagram of diagnostic apparatus of embodiment The figure which shows typically normal / abnormal discrimination by SVM Diagram showing normal clusters used in cluster analysis

Explanation of symbols

2 Overhead traveling vehicle system 4 Interbay route 6 Intrabay route 8 Maintenance beirut route 10 Round route 12 Branching junction 14 Curve 16 Overhead traveling vehicle 18 Station 20 Buffer 22 Maintenance area 24 Overhead traveling vehicle controller 26 Diagnostic device 30 Traveling rail 31 Upper rail 32 Lower portion Rail 33 Litz wire 34 Traveling carriage 35 Driving wheel 36 Driven wheel 37 Guide wheel 38 Traveling motor 39 Vibration sensor 40 Power receiving unit 41 Frame 42 Lateral feeding unit 43 Lateral feeding motor 44 Turntable 45 Rotating motor 46 Elevating drive unit 47 Elevating motor 48 Lift platform 49 Cover 50 Article 51 Travel speed pattern 52 Acceleration section 53 Constant speed section 54 Deceleration section 55 Elevation speed pattern 56 Lower section 57 Up section 58 Lateral feed speed pattern 59 Advance section 60 Return section 6 Input interface 64 storage unit 66 statistically unit 68 SVM (support vector machine) group 70 cluster analysis unit 72 GUI (Graphical User Interface)
74 Maintenance management unit 76 Teacher data input unit

Claims (5)

A moving body diagnostic system that moves according to a predetermined speed pattern consisting of acceleration, constant speed movement and deceleration along a finite and predetermined movement path,
A sensor for detecting vibrations of the moving body, and a monitoring means for monitoring the state of the moving motor of the moving body;
From the sensor output and the output of the monitoring means, a plurality of feature quantities indicating the process from the start to the stop of the moving body, or the characteristics of each section divided into an acceleration area, a constant speed area, and a deceleration area are obtained. A feature amount calculating means for
A moving body diagnosis comprising: a plurality of obtained feature amounts as vectors in a multidimensional vector space; and a determination unit for determining normality / abnormality of the moving body from a position in the vector space. system. 2. The moving body diagnosis system according to claim 1, wherein a plurality of the discriminating means are provided corresponding to the type of abnormality of the moving body. Means for storing a cluster consisting only of a vector obtained from a normal moving body, and for determining whether a vector consisting of a plurality of feature values inputted from the moving body belongs to the cluster The mobile body diagnosis system according to claim 1, further comprising a cluster analysis unit. 6. A tracked carriage that travels along a traveling rail and moves up and down an article, wherein at least normality / abnormality is diagnosed with respect to running and lifting of the tracked carriage. The diagnostic system of the mobile body in any one of 1-3. With a plurality of overhead traveling vehicles where the moving body travels on the traveling rail provided in the ceiling space in the clean room,
The overhead traveling vehicle is provided with a vibration detection sensor and a monitoring means for monitoring the state of the traveling motor and the lifting motor,
In the diagnostic device, normality / abnormality of the overhead traveling vehicle is determined based on a vector composed of the feature amount of the signal from the sensor and the monitoring means, and other than the traveling rail for the overhead traveling vehicle determined to be abnormal. 5. The mobile body diagnosis system according to claim 4, wherein re-diagnosis is performed from the data in the section.

Images