JP6139149B2 - Analysis method and analysis apparatus - Google Patents

Description

  The present invention relates to an analysis method and an analysis apparatus.

  2. Description of the Related Art There is known a technique for recording data related to the operational state of an aircraft and analyzing the soundness (safety) of the aircraft based on the recorded data. For example, Patent Document 1 discloses recording a period during which the display value of an indicator exceeds a threshold value and the degree thereof during operation of the aircraft.

JP-A-3-129600

  As described above, when the analysis is performed using only the data selected based on the threshold value, only a part of the data related to the assumed abnormality is used. is there. The present invention has been made in view of the above, and an object of the present invention is to provide an analysis method and an analysis apparatus that can analyze an abnormality that is not assumed in advance.

  In one aspect, an analysis method for detecting an anomaly in an aircraft includes obtaining raw data related to a state of operation of the aircraft, and applying multivariate analysis to the raw data to detect the anomaly in the aircraft Detecting.

  In another aspect, an analysis apparatus for detecting an abnormality of an aircraft includes a data acquisition unit that acquires raw data related to a state of operation of the aircraft, and the aircraft by applying multivariate analysis to the raw data. And a multivariate analysis unit for detecting abnormalities.

  The analysis method and the analysis apparatus according to the present invention have an effect of being able to analyze an abnormality that is not assumed in advance.

FIG. 1 is a diagram illustrating a configuration of an analysis system according to an embodiment. FIG. 2 is a diagram showing the configuration of the analysis program. FIG. 3 is a flowchart showing an example of the operation of the storage unit of the aircraft. FIG. 4 is a flowchart illustrating an example of analysis processing. FIG. 5 is a flowchart illustrating another example of the analysis process. FIG. 6 is a flowchart illustrating an example of abnormality detection processing based on a threshold. FIG. 7 is a flowchart illustrating an example of abnormality detection processing based on multivariate analysis. FIG. 8 is a flowchart illustrating an example in which an abnormality detection process based on a threshold and an abnormality detection process based on multivariate analysis are executed in association with each other.

  Hereinafter, embodiments of an analysis method and an analysis apparatus will be described in detail with reference to the drawings. Note that the present invention is not limited to the embodiments. In addition, the constituent elements in this embodiment include those that can be easily assumed by those skilled in the art, those that are substantially the same, and those in a so-called equivalent range.

  A configuration of the analysis system according to the present embodiment will be described. FIG. 1 is a diagram illustrating a configuration of an analysis system according to the present embodiment. The analysis system shown in FIG. 1 includes aircrafts 10a to 10x, an analysis device 20, a maintenance management system 30, and a parts ordering system 40. In the following description, the aircrafts 10a to 10x may be collectively referred to as the aircraft 10 without specifying which one.

  The aircraft 10 includes, for example, a fixed wing aircraft and a rotary wing aircraft. The aircraft 10 includes a control unit 11, detectors 12a to 12m, devices 13a to 13n, and a storage unit 14. In the following description, the detectors 12a to 12m are collectively referred to as the detector 12 without specifying which one, and the devices 13a to 13n are collectively referred to as the device 13 without specifying any. Sometimes.

  The control unit 11 performs various controls related to the operation of the aircraft 10. For example, the control unit 11 determines how to control the device 13 based on the operation received by the operation unit or the value detected by the detector 12, and the control signal corresponding to the determination corresponds to the device. 13 to send.

  The detector 12 detects a value related to the situation of the corresponding part and outputs the detected value. The detector 12 detects, for example, acceleration indicating the magnitude and period of vibration, acceleration indicating the magnitude of an applied external force, temperature, rotation speed, and the like. The detector 12 includes one used for control by the control unit 11, one used for analysis by the analysis system according to the present embodiment, and one used for both.

  The equipment 13 is a unit that physically constitutes an aircraft for the operation of the aircraft. The device 13 includes, for example, a unit that functions as an operation unit that receives an operation. Such a unit transmits a control signal for controlling another device 13. In addition, the equipment 13 includes units that are driven to fly the aircraft 10, such as, for example, engines, transmissions, and flaps. Such a unit operates based on the transmitted control signal.

  The storage unit 14 includes a storage medium such as a solid state drive or a hard disk drive. The storage unit 14 receives the value output from the detector 12, the control signal transmitted from the control unit 11 to the device 13, and the control signal transmitted between the devices 13, for example, by snooping the bus. Then, the storage unit 14 records the received information as raw data in the monitoring data 14a in association with information indicating the type of information and time. Raw data is data that has not been thinned out by threshold processing or the like. That is, the storage unit 14 records not only information that is determined to be abnormal based on a known standard, but also information that is determined to be normal based on a known standard.

  The raw data does not include incidental information such as information for synchronization of communication unless substantial information on the operational status of the aircraft such as detected values and controlled variables is lost. Also good. The raw data may be subjected to processing for reducing the storage capacity, such as compression processing, as long as substantial information regarding the operational state of the aircraft such as the detected value and control amount is not lost.

  The analysis device 20 performs analysis related to the soundness (safety) of the aircraft 10. The analysis device 20 includes a control unit 21, an operation unit 22, a display unit 23, a communication unit 24, and a storage unit 25.

  The control unit 21 comprehensively controls the analysis device 20. Specifically, the control unit 21 includes an arithmetic device such as a CPU (Central Processing Unit) and a storage device such as a RAM (Random Access Memory). And the control part 21 implement | achieves various functions by expand | deploying the program memorize | stored in the memory | storage part 25 to the storage area allocated on the memory | storage device, and performing with an arithmetic unit.

  The operation unit 22 receives an operation. The display unit 23 displays various information such as information for using the analysis device 20 and analysis results of the analysis device 20. The communication unit 24 realizes communication with other devices (systems) such as the maintenance management system 30 and the parts ordering system 40.

  The storage unit 25 includes a non-transitory storage medium such as a solid state drive or a hard disk drive. The storage unit 25 stores data and programs. The data stored in the storage unit 25 includes monitoring data 25a, threshold data 25b, and abnormality data 25c. The program stored in the storage unit 25 includes an analysis program 25d.

  The storage unit 25 may include a combination of a portable non-transitory storage medium such as a memory card, an optical disk, or a magneto-optical disk, and a storage medium reader. In this case, some or all of the data and programs stored in the storage unit 25 in FIG. 1 may be stored in a portable storage medium. In FIG. 1, some or all of the data and programs stored in the storage unit 25 may be downloaded from other devices through communication by the communication unit 24.

  The monitoring data 25a is data in which the monitoring data 14a recorded in the aircrafts 10a to 10x is accumulated. The monitoring data 14a is transmitted to the analysis device 20 via some medium and added to the monitoring data 25a. The medium used for transmitting the monitoring data 14a may be a portable non-transitory storage medium, a physical communication line, or a wireless communication line. The monitoring data 14 a may be transmitted to the analysis device 20 in real time almost simultaneously with being recorded in the storage unit 14, or may be transmitted to the analysis device 20 after being accumulated in the storage unit 14. Data is stored in the monitoring data 25a so that it can be determined with which aircraft the data is acquired.

  The threshold data 25b stores a threshold for detecting an abnormality by threshold processing. Information relating to the detected abnormality is stored in the abnormality data 25c. The information relating to the detected abnormality includes information relating to the details of the abnormality, information for identifying the part where the abnormality is detected, and information for identifying the aircraft 10 including the part.

  The analysis program 25d provides a function of analyzing an abnormality based on the monitoring data 25a. FIG. 2 is a diagram showing the configuration of the analysis program 25d. As shown in FIG. 2, the analysis program 25 d includes a data acquisition unit 26, a threshold processing unit 27, and a multivariate analysis unit 28.

  The data acquisition unit 26 provides a function of accumulating the monitoring data 14a recorded in the aircrafts 10a to 10x as the monitoring data 25a.

  The threshold processing unit 27 provides a function of detecting a known abnormality. Specifically, the threshold processing unit 27 provides a function of extracting data indicating abnormality from the monitoring data 25a by threshold processing using the threshold stored in the threshold data 25b, and detecting abnormality based on the extracted data. To do.

  The multivariate analysis unit 28 provides a function of detecting an abnormality that is not assumed in advance. Specifically, the multivariate analysis unit 28 provides a function of detecting an abnormality by applying multivariate analysis to the monitoring data 25a. More specifically, the multivariate analysis unit 28 provides a function of detecting an outlier by multivariate analysis and detecting an abnormality based on the detected outlier. For example, a Mahalanobis Taguchi system can be used to detect outliers. A technique for detecting an abnormality using the Mahalanobis Taguchi system is disclosed in, for example, Japanese Patent Application Laid-Open No. 2012-159298.

  In the multivariate analysis by the function provided by the multivariate analysis unit 28, anomalies are detected by detecting outliers using a statistical method. For this reason, it is possible to detect an abnormality that is not assumed in advance (an unknown abnormality in which a threshold is not set in advance). Abnormalities that are not assumed in advance include, for example, abnormalities that do not cause a failure immediately but may cause a failure if left unattended.

  Furthermore, in the multivariate analysis by the function provided by the multivariate analysis unit 28, data that has not been extracted as data indicating abnormality in the threshold processing by the function provided by the threshold processing unit 27 is also used. For this reason, according to the function provided by the multivariate analysis unit 28, it is possible to detect an abnormality that cannot be detected by analyzing the data extracted by the threshold processing.

  The maintenance management system 30 manages the maintenance work of the aircrafts 10a to 10x. Based on the determination result of the analysis device 20, the maintenance management system 30 performs measures such as replacing the part where the abnormality is detected or the timing of inspection. The maintenance management system 30 may perform a procedure for increasing the inspection items in the periodic inspection of the aircrafts 10a to 10x based on the determination result of the analysis device 20. By such a measure, the possibility of occurrence of a failure due to an abnormality that is not assumed in advance can be reduced, and the soundness (safety) of the aircraft 10a to 10x can be increased.

  The parts ordering system 40 manages the ordering of parts of the aircrafts 10a to 10x. Based on the determination result of the analysis device 20, the part ordering system 40 performs a procedure such as ordering a part at a site where an abnormality has been detected. The parts ordering system 40 may place an order only for parts whose inventory is less than a threshold or parts whose lead time to delivery is longer than a threshold among the parts at which an abnormality is detected. Thus, by ordering parts early, it is possible to reduce the possibility that a situation in which the aircraft 10a to 10x cannot be operated due to waiting for delivery of replacement parts will occur. Furthermore, since unnecessary parts do not need to be stocked excessively, the part management cost can be reduced.

  The operation of the storage unit 14 of the aircraft 10 will be described with reference to FIG. FIG. 3 is a flowchart illustrating an example of the operation of the storage unit 14. The operation shown in FIG. 3 is started when the aircraft 10 is powered on.

  The storage unit 14 determines whether a control signal for the device 13 has been received (step S101). When the control signal is received (step S101, Yes), the storage unit 14 records the control signal in the monitoring data 14a in association with the time or the like (step S102). When the control signal is not received (step S101, No), the storage unit 14 does not execute step S102.

  Subsequently, the storage unit 14 determines whether the detection value of the detector 12 has been received (step S103). When the detected value is received (step S103, Yes), the storage unit 14 records the detected value in the monitoring data 14a in association with the time or the like (step S104). When the detected value is not received (No at Step S103), the storage unit 14 does not execute Step S104.

  Subsequently, the storage unit 14 determines whether to end the operation (step S105). For example, the storage unit 14 determines to end the operation when the aircraft is powered off. When it determines with complete | finishing operation | movement (step S105, Yes), the memory | storage part 14 complete | finishes the operation | movement shown in FIG. When it is not determined to end the operation (No at Step S105), the storage unit 14 returns to Step S101. In this way, the aircraft 10 records raw data related to the state of operation. Although FIG. 3 shows an example in which the process for recording the control signal and the process for recording the detection value are executed in series, the execution form of these processes is not limited to this. These processes may be executed in parallel. Further, the execution order of these processes may be reverse to the order shown in FIG.

  The operation of the analysis apparatus 20 will be described with reference to FIGS. FIG. 4 is a flowchart illustrating an example of analysis processing executed by the analysis apparatus 20. The analysis process shown in FIG. 4 is realized by the control unit 21 executing the analysis program 25d.

  The control unit 21 performs an abnormality detection process based on the threshold value (step S201). The abnormality detection process based on the threshold is a detection process using a function provided by the threshold processing unit 27, that is, a process of detecting an abnormality based on a known standard. In parallel with this, the control unit 21 performs abnormality detection processing based on multivariate analysis (step S202). The abnormality detection process based on the multivariate analysis is a detection process by a function provided by the multivariate analysis unit 28, that is, a process of detecting an abnormality that is not assumed in advance.

  FIG. 4 shows an example in which the abnormality detection process based on the threshold and the abnormality detection process based on the multivariate analysis are executed in parallel, but the execution form of these processes is not limited to this. For example, as illustrated in FIG. 5, the control unit 21 may perform an abnormality detection process based on a threshold (step S201) and then perform an abnormality detection process based on multivariate analysis (step S202). Good. The abnormality detection process based on multivariate analysis may be performed before the abnormality detection process based on the threshold value.

  The execution frequency of the abnormality detection process based on the multivariate analysis may be different from the execution frequency of the abnormality detection process based on the threshold value. In the abnormality detection process based on the threshold value, an abnormality indicating a failure sign is detected with high accuracy. Further, the abnormality detection process based on the threshold value requires a relatively short time because the calculation amount is small. For this reason, the abnormality detection process based on the threshold value may be executed with a relatively high frequency. For example, the abnormality detection process based on the threshold value may be executed every time the operation of the aircraft 10 is completed.

  On the other hand, the abnormality detection process based on multivariate analysis requires a relatively long time because of the large amount of calculation. Furthermore, a highly urgent abnormality can be detected with high accuracy by an abnormality detection process based on a threshold value. For this reason, the abnormality detection process based on multivariate analysis may be executed at a relatively low frequency. For example, the abnormality detection process based on multivariate analysis may be executed once every several days or once a week.

  FIG. 6 is a flowchart illustrating an example of abnormality detection processing based on the threshold values illustrated in FIGS. 4 and 5. The control unit 21 acquires data that has not been determined by the threshold value from the monitoring data 25a (step S301). When the data can be acquired (step S302, Yes), the control unit 21 acquires a threshold corresponding to the data from the threshold data 25b, and compares the detection value or the control amount included in the data with the threshold (step S303).

  When it is determined by the comparison that there is an abnormality (step S304, Yes), the control unit 21 adds information related to the abnormality to the abnormality data 25c (step S305). Thereafter, the control unit 21 returns to Step S301.

  When it is not determined by comparison that there is an abnormality (No at Step S304), the control unit 21 returns to Step S301 without executing Step S305.

  When the data that has not been determined by the threshold value cannot be acquired from the monitoring data 25a, that is, when the determination of all data has been completed (No in step S302), the control unit 21 ends the abnormality detection process based on the threshold value. To do.

  FIG. 7 is a flowchart illustrating an example of an abnormality detection process based on the multivariate analysis illustrated in FIGS. 4 and 5. The control unit 21 selects one of a plurality of preset ranges, and acquires data of the selected range from the monitoring data 25a in a form in which related data is associated based on time (step) S401). In step S401, data is acquired regardless of whether or not there is an abnormality in the abnormality detection process based on the threshold value.

  The range defines which aircraft data should be acquired, which part data should be acquired, what type of data should be acquired, which period of data should be acquired, and so on. A plurality of ranges are set so that outliers can be detected from various viewpoints. For example, the certain range is set so that data related to one part of one aircraft is acquired for a certain period of time. The certain range is set such that data related to one part of a plurality of aircraft operated in the same form is acquired for the period operated in the same form. The range may be set so that one data is included in a plurality of ranges.

  Acquiring data in a form in which related data is associated with each other based on time means that data related to the same action is acquired in combination. For example, the control unit 21 acquires the data of the selected range in a form in which the data of the control amount and the data of the detection value at the same time related to the part controlled by the control amount are combined. For example, the control unit 21 acquires data in a selected range in a form in which data of a plurality of types of detection values at the same time regarding the same part of the same aircraft are combined. For example, the control unit 21 acquires data in a selected range in a form in which data of detection values at the same time regarding a plurality of related parts of the same aircraft are combined. By acquiring data in this way, it becomes possible to perform more accurate analysis.

  The process of associating related data based on time may be executed in advance. For example, the storage unit 14 may execute processing for associating related data based on time, and store the data in the monitoring data 14a in a correlated form. For example, the control unit 21 executes a process of associating related data based on time based on the function provided by the data acquisition unit 26, and stores the data in the associated form in the monitoring data 25a. Good. Thus, by associating the data in advance, the time required for the abnormality detection process based on multivariate analysis can be shortened.

  The control unit 21 executes multivariate analysis processing using the data acquired in step S401 (step S402). The details of the multivariate analysis process may differ depending on the range selected in step S401. And the control part 21 determines whether the outlier was detected by the multivariate analysis process (step S403).

  When an outlier is detected (step S403, Yes), the control unit 21 adds information regarding the abnormality indicated by the outlier to the abnormality data 25c (step S404). When an outlier is not detected (step S403, No), the control unit 21 does not execute step S404.

  Subsequently, the control unit 21 determines whether or not the processing for all ranges has been completed (step S405). When the processing of the entire range has not been completed (step S405, No), the control unit 21 returns to step S401. When the process of all the ranges is completed (step S405, Yes), the control part 21 complete | finishes the abnormality detection process based on multivariate analysis.

  In addition, the aspect of this invention shown by said Example can be arbitrarily changed in the range which does not deviate from the summary of this invention. For example, the program shown in the above embodiment may be divided into a plurality of modules or may be integrated with other programs. The data shown in the above embodiment may be divided into a plurality of data or may be integrated with other data. The analysis device 20 may be configured to be installed in the aircraft 10 and notify the pilot of data stored in the abnormality data 25c.

  In the above embodiment, the example in which the abnormality detection process based on the threshold and the abnormality detection process based on the multivariate analysis are executed independently has been described. However, these processes may be executed in association with each other. FIG. 8 is a flowchart illustrating an example in which an abnormality detection process based on a threshold and an abnormality detection process based on multivariate analysis are executed in association with each other. In the example shown in FIG. 8, the threshold data update process is executed (step S203) after the execution of the abnormality detection process based on multivariate analysis (step S202).

  In the threshold data update process, the control unit 21 analyzes the correlation between the outlier detected in the course of the abnormality detection process based on multivariate analysis and the abnormality detected in the abnormality detection process based on the threshold. When it is determined that there is a correlation higher than a certain value between the outlier and the abnormality detected in the abnormality detection process based on the threshold, the control unit 21 sets a new threshold based on the outlier as threshold data 25b. Add to The new threshold value is set so as to deviate sufficiently from the normal value. The function of setting a new threshold is implemented, for example, by providing a threshold setting unit in the analysis program 25d.

  In this way, by setting a new threshold based on the outlier, it is possible to increase the possibility that a sign of failure can be detected at an early stage. Further, in the case of an aircraft, in order to avoid the occurrence of a serious accident, countermeasures such as replacement are often taken for the corresponding part when a sign of failure is detected. For this reason, it is difficult to provide a new threshold for detecting an abnormality based on the track record of failure. By setting a new threshold value based on the outlier, it is possible to newly set a threshold value having high detection accuracy of a failure without being based on the actual track record of the failure.

  When a new threshold value based on the outlier is added to the threshold data 25b (step S204, Yes), the control unit 21 executes an abnormality detection process based on the added threshold value (step S205). In the abnormality detection process based on the added threshold value, data already compared with another threshold value in the abnormality detection process based on the threshold value is also compared with the added threshold value. Then, analysis is performed on data determined to have an abnormality based on the added threshold, and information regarding the abnormality is added to the abnormality data 25c.

  The analysis device 20 may be configured to calculate the cost required for maintenance of the aircraft 10 based on the abnormality data 25c. As described above, the abnormality data 25c stores information related to the abnormality of the aircraft 10, including an abnormality that cannot be detected by a known standard. By using these pieces of information, it is possible to calculate the cost required for inspection, parts replacement, etc. to avoid a failure with high accuracy.

  Analysis device 20 may be configured to optimize a schedule for maintenance of aircraft 10 based on abnormality data 25c. The analysis device 20 can detect a sign of a failure early by detecting an abnormality that is not assumed in advance. Thereby, the range of selection of the time for performing maintenance work such as inspection and parts replacement is widened. For this reason, for example, the analysis device 20 creates a schedule so that maintenance work is performed collectively for each aircraft 10 within a certain period, thereby ensuring soundness (safety) and The operational time can be increased.

10a to 10x Aircraft 11 Control units 12a to 12m Detectors 13a to 13n Equipment 14 Storage unit 14a Monitoring data 20 Analyzing device 21 Control unit 22 Operation unit 23 Display unit 24 Communication unit 25 Storage unit 25a Monitoring data 25b Threshold data 25c Abnormal data 25d Analysis program 26 Data acquisition unit 27 Threshold processing unit 28 Multivariate analysis unit 30 Maintenance management system 40 Parts ordering system

Claims (14)

An analysis method for detecting an abnormality in an aircraft,
Obtaining raw data relating to the operational status of the aircraft;
Detecting an abnormality of the aircraft by applying multivariate analysis to the raw data ,
The step of acquiring the raw data sets and sets a range so that data related to one part of the plurality of aircraft operated in the same form is acquired for a period operated in the same form Data in the selected range,
The step of detecting an abnormality of the aircraft is an analysis method of detecting an abnormality of the aircraft by applying multivariate analysis to the raw data in the acquired range . Calculating a cost required for maintenance of the aircraft based on the detected abnormality data of the aircraft;
The analysis method according to claim 1, further comprising: Optimizing a schedule for maintenance of the aircraft based on the detected abnormality data of the aircraft;
The analysis method according to claim 1 or 2, further comprising: The analysis method according to claim 1, further comprising a step of detecting an abnormality of the aircraft by applying a threshold process to the raw data. In the step of detecting an abnormality of the aircraft by applying a multivariate analysis the raw data, according to claim 4, also on the determined data that there is no abnormality by said threshold processing said multivariate analysis is applied analysis method. The multivariate based on outliers detected by analysis, the method of analysis according to claim 4 or 5 further comprising the step of setting a threshold used in the threshold process. In the step of obtaining the raw data, method of analysis according to any one of claims 1 to 6, relevant data is acquired in association on the basis of time. An analysis device for detecting abnormalities in an aircraft,
A data acquisition unit for acquiring raw data related to the state of operation of the aircraft;
A multivariate analysis unit that detects an abnormality of the aircraft by applying multivariate analysis to the raw data ,
The data acquisition unit sets a range such that data related to one part of the plurality of aircraft operated in the same form is acquired for a period operated in the same form, Get the data,
The multivariate analysis unit detects an abnormality of the aircraft by applying multivariate analysis to the raw data in the acquired range . The analysis apparatus according to claim 8, wherein a cost required for maintenance of the aircraft is calculated based on the detected abnormality data of the aircraft. The analysis apparatus according to claim 8 or 9, wherein a schedule related to maintenance of the aircraft is optimized based on the detected abnormality data of the aircraft. The analysis device according to claim 8 , further comprising a threshold processing unit that detects an abnormality of the aircraft by applying threshold processing to the raw data. The analysis device according to claim 11 , wherein the multivariate analysis unit applies the multivariate analysis to data determined to be normal by the threshold processing unit. The analysis device according to claim 11 or 12 , further comprising a threshold setting unit configured to set a threshold used in the threshold processing unit based on an outlier detected by the multivariate analysis by the multivariate analysis unit. The data acquisition unit, the analysis apparatus according to any one of claims 8 to 13 for obtaining association based on the time the associated data.

Images