DATA MANAGEMENT DEVICE, DATA MANAGEMENT PROGRAM, AND DATA MANAGEMENT PROCEDURES

Abstract

Data management device (1) that can calculate a rating index based on sensor data, the data management device (1) comprising: a rating index calculation unit (102) that calculates the rating index using the sensor data obtained from one or more sensors (60) installed in an industrial machine; a sensor data storage unit (103) that stores the sensor data; and a sensor data erasure unit (104) that erases the sensor data from the sensor data storage unit (103) when the total amount of sensor data in the sensor data storage unit (103) exceeds an upper limit, wherein the sensor data erasure unit (104) determines a storage priority of the sensor data based on a degree of change of the rating index and a storage density of the sensor data and erases the sensor data according to the storage priority.

Description

GENERAL STATE OF THE ART Field of invention

[0001] The present invention relates to a data management device, a data management program, and a data management method, and in particular relates to a data management device, a data management program, and a data management method for determining a storage method for the sensor data by determining a priority of sensor data. Description of the state of the art

[0002] Techniques for detecting anomalies in an industrial machine (hereinafter referred to as the machine), such as a machine tool, using sensor data from sensors installed in the machine, are well known. The sensor data obtained from the sensors can be used after processing, rather than directly. For example, there is a device that converts time-series values ​​of the sensor data into an index expressing the degree of anomaly in the machine (hereinafter referred to as the anomaly level) and displays the anomaly level to a user (hereinafter referred to as the data management device). Based on the anomaly level displayed by the data management device, the user makes a determination regarding the need to replace components or the like.

[0003] The user can then refer to the sensor data that formed the basis for calculating the degree of anomaly and use this data as a reference for the determination. Accordingly, conventional data management devices do not delete the sensor data even after calculating the degree of anomaly, but rather store it. However, this presents a problem, as storing all sensor data can lead to an enormous amount of data, given the generally long lifespan of the machine.

[0004] International Publication No. WO 2016 / 199210 discloses a system that calculates an anomaly level based on sensor data, calculates an anomaly frequency and remaining lifetime based on the anomaly level, and determines which sensor data to store based on the anomaly frequency and remaining lifetime, and the like. Japanese Patent Publication No. 2017-173321 discloses a system that calculates an anomaly level based on sensor data and then deletes the sensor data according to a reference such as the age of the measurement time.

[0005] The technology disclosed in International Publication No. WO 2016 / 199210, however, requires the ability to calculate the anomaly frequency and remaining lifetime. This calculation necessitates prior knowledge of anomalies and failures of the device. In contrast, the technology disclosed in Japanese Patent Publication No. 2017-173321 has no upper limit on the amount of sensor data that can be stored, which can lead to insufficient storage space. Even if sufficient free space is available in the storage area, the data may be unnecessarily thinned out.

[0006] The present invention was developed to solve these problems. One object of the invention is to provide a data management device, a data management program, and a data management method for determining a storage method for sensor data by determining the priority of sensor data. BRIEF SUMMARY OF THE INVENTION

[0007] A data management device according to one embodiment of the invention is a data management device that can calculate a rating index based on sensor data. The data management device comprises a rating index calculation unit that calculates the rating index using sensor data obtained from one or more sensors installed in an industrial machine; a sensor data storage unit that stores the sensor data; and a sensor data erasure unit that erases the sensor data from the sensor data storage unit when the total amount of sensor data in the sensor data storage unit exceeds an upper limit. The data management unit is characterized in that the sensor data erasure unit determines a storage priority for the sensor data based on the degree of change in the rating index and erases the sensor data according to this storage priority.

[0008] The sensor data erasure unit of the data management device according to an embodiment of the invention is characterized in that it uses a dispersion of the rating index or a time derivative of the rating index in a prescribed period as the degree of change of the rating index.

[0009] The sensor data erasure unit of the data management device according to an embodiment of the invention is characterized in that it determines the storage priority of the sensor data based on the degree of change of the rating index and the time that has elapsed since the sensor data was obtained.

[0010] The sensor data erasure unit of the data management device according to an embodiment of the invention is characterized in that it determines the storage priority of the sensor data based on the degree of change of the rating index and a storage density of the sensor data.

[0011] The data management device according to one embodiment of the invention further comprises a user interface unit that displays time-series data of the rating index, displays a selection of the rating index at a specific time, and displays time-series data of the sensor data that formed a basis for calculating the selected rating index. The data management device is characterized in that, when the sensor data that formed the basis for calculating the selected rating index has been deleted, the user interface unit displays other sensor data that was obtained at a time closest to the time the deleted sensor data was obtained.

[0012] The data management device according to one embodiment of the invention is characterized in that the calculated rating index is an anomaly level.

[0013] A data management program according to one embodiment of the invention is a data management program that causes a computer to execute a process of calculating a rating index based on sensor data. The data management program causes the computer to perform a first step of calculating the rating index using the sensor data obtained from one or more sensors installed in an industrial machine; a second step of storing the sensor data in a sensor data storage unit; and a third step of deleting the sensor data from the sensor data storage unit when the total amount of sensor data in the sensor data storage unit exceeds an upper limit.The data management program is characterized by the fact that in the third step, a storage priority of the sensor data is determined based on a degree of change in the rating index, and the sensor data is deleted according to the storage priority.

[0014] A data management method according to one embodiment of the invention is a data management method for calculating a rating index based on sensor data. The data management method performs a first step of calculating the rating index using sensor data obtained from one or more sensors installed in an industrial machine; a second step of storing the sensor data in a sensor data storage unit; and a third step of deleting the sensor data from the sensor data storage unit when the total amount of sensor data in the sensor data storage unit exceeds an upper limit. The data management method is characterized in that, in the third step, a storage priority for the sensor data is determined based on the degree of change in the rating index, and the sensor data is deleted according to this storage priority.

[0015] According to the invention, a data management device, a data management program, and a data management method can be provided to determine a storage method for the sensor data by determining the priority of sensor data. List of characters

[0016] The above-mentioned and other tasks and features of the invention will become apparent from the following description of an embodiment with reference to the accompanying drawings, wherein Fig. 1 is a diagram that shows an example of a hardware layout of a data management device; Fig. 2 is a diagram that illustrates an example of a functional layout of the data management device; Fig. 3 is a diagram that provides an example of anomaly levels; Fig. 4 is a diagram that represents a display example of sensor data; Fig. 5 is a diagram that represents a relationship between a degree of change in the anomaly level and a storage priority; and Fig. 6 is a flowchart that provides an example of how the data management device operates. DETAILED DESCRIPTION OF THE PREFERRED EXECUTION FORM

[0017] Fig. Figure 1 is a schematic hardware setup representing a data management device. 1 represents the data management device. 1 A data management device is an information processing device that performs various processes using sensor data acquired by a machine. 1 Examples include a personal computer (PC), a numerical control unit, or similar devices. The data management unit 1 a CPU 11 , a ROM 12 , a RAM 13 , a non-volatile memory 14, an input / output device 15 , an interface 16 and a bus 10 up. To the data management device 1 are one or more sensors 60 connected.

[0018] The CPU 11 is a processor that manages the data management device 1 generally controls the CPU. 11 reads system programs located in the ROM 12 are stored via the bus 10 and controls the entire data management device 1 according to the system programs.

[0019] The system programs were pre-programmed into the ROM. 12 saved.

[0020] The RAM 1 stores temporary calculation data, display data, or data, programs, or the like, which are entered by an operator via the input / output device 15 to be entered temporarily.

[0021] The non-volatile memory 14For example, it is supported by a battery (not shown) and retains its memory state even when the data management device is down. 1 is switched off. The non-volatile memory 14 For example, it stores data, programs, or the like, which are transmitted through the input / output device. 15 The programs and data stored in non-volatile memory are entered. 14 are stored, can be accessed in RAM during execution and use 13 will be loaded.

[0022] The input / output device 15 A data input / output device is a device that includes a display device, such as a screen, and an input device, such as a keyboard. The input / output device 15 For example, it displays information that is processed by the CPU. 11 received, on the screen. The input / output device 15Sends data entered via the keyboard to the CPU. 11 further.

[0023] The sensors 60 These are sensors that are installed at various points in the machine, such as temperature sensors, speed sensors, or acceleration sensors. One or more of these sensors 60 is or are connected to the interface via a wired or wireless communication device 16 connected. The sensor data generated by the sensors 60 output will be via the interface 16 to the CPU 11 passed on.

[0024] Fig. Figure 2 is a block diagram showing a schematic functional structure of the data management device. 1 represents the data management device. 1 It features a sensor data acquisition unit 101 , a valuation index calculation unit 102 , a sensor data storage unit103 , a sensor data erasure unit 104 , a rating index storage unit 105 , and a user interface unit 106 on.

[0025] The sensor data acquisition unit 101 The sensor data is obtained, for example, through communication at regular intervals with one or more sensors. 60 , which is or are installed in the machine. The sensor data acquisition unit 101 The sensor data acquired is collected sequentially in the sensor data storage unit. 103 That is, the sensor data acquisition unit 101 forms in the sensor data storage unit 103 Time series sensor data.

[0026] The valuation index calculation unit 102 calculated using the data stored in the sensor data storage unit 103A rating index is assigned to the collected sensor data. This rating index can be, for example, an anomaly grade. The anomaly grade is an index calculated based on the time-series sensor data and expresses the degree of an anomaly in the machine. Although there are many types of anomaly grades, the methods for calculating these different types are omitted here, as they are generally known. The rating index calculation unit 102 For example, using multiple sensor data elements over a given period (usually time series data on the order of several kilobytes), it calculates an anomaly degree data element. That is, the assessment index calculation unit. 102 Based on the time series sensor data, it generates time series anomaly degree data with a smaller data size.

[0027] The sensor data storage unit 103The data acquired by the sensor unit is stored 101 Time series sensor data obtained in a specific storage area.

[0028] If a total set of sensor data stored in the sensor data storage unit 103 If the stored data exceeds a predetermined upper limit, the sensor data erasure unit deletes it. 104 The sensor data is deleted in ascending order of storage priority, determined by a technique described subsequently, until the total amount is equal to or less than the upper limit. That is, the sensor data deletion unit... 104 In this embodiment, the acquired sensor data is stored without modification until the need arises for deletion.

[0029] The sensor data erasure unit 104The storage priority of the sensor data is determined based on at least one of the following: the degree of change of the rating index calculated by the rating index calculation unit (such as the anomaly level), the storage density of the sensor data, and the time elapsed since the sensor data was acquired. Here, the degree of change of the rating index refers to the extent to which the rating index changes. The storage density of the sensor data refers to the number of sensor data points stored per unit of time. A relationship between these factors is represented by expression (1). P r ( x ) = f ( d e n ( x ) , d i f ( x ) , E l a ( x ) )

[0030] Pr(x) is the storage priority of the sensor data acquired at a measurement point (i.e., a measurement time) x. den(x) is the storage density of the sensor data in the vicinity of measurement point x (for example, the number of data points stored in ±12 hours). dif(x) is the degree of change of the rating index in the vicinity of measurement point x (for example, a variation in the anomaly level or a time derivative of the anomaly level in ±12 hours). Ela(x) is the time elapsed from measurement point x to the current time.

[0031] Typically, it is not the sensor data itself, but rather a highly recognizable rating index (such as a curve indicating a temporal transition in the degree of anomaly) that the user refers to in order to determine the presence or absence of an anomaly in the machine. However, the user may, if necessary, refer to the sensor data that formed the basis for calculating the rating index. Traditionally, for example, a memory area with a huge capacity was used to store all the sensor data. In this regard, the inventor found that a large number of users, considering a rating index that is of high importance (attracts user interest), are strongly inclined to refer to the sensor data that formed the basis for calculating the rating index.The rating index, which is of high importance (attracting user interest), refers, for example, to a rating index around a time when the rating index fluctuates significantly, that is, when the rating index shows a high degree of change. Based on this finding, the sensor data deletion unit increases 104 In this embodiment, the storage priority of sensor data is set at locations where the anomaly level shows a high degree of change, in order to ensure that the sensor data resists deletion. At locations where the anomaly level shows a low degree of change, the sensor data deletion unit reduces the storage priority. 104 the storage priority of the sensor data to ensure that the sensor data becomes vulnerable to deletion (see Fig. 5) This makes it possible to preferentially delete sensor data where the likelihood of the user referring to it is low.

[0032] However, in a case where all sensor data is deleted at points where the anomaly level shows a low degree of change, it becomes impossible to refer back to the sensor data when necessary, even though there is a small probability that the user might refer to it. It is desirable that, for at least some anomaly level data points, the sensor data that formed the basis for calculating the anomaly level be stored during periods when the anomaly level shows a low degree of change. Based on this understanding, the sensor data deletion unit increases 104In this embodiment, the storage priority of the sensor data increases the lower the storage density of the sensor data (the number of stored data points) in the vicinity of the measurement point, in order to ensure that the sensor data resists deletion. Thus, an appropriate thinning rate for the sensor data can be maintained at points where the anomaly level indicates a low degree of change, so that the necessary sensor data can be referenced.

[0033] Furthermore, the inventor discovered that the frequency of references to sensor data increases for a newer rating index, while the frequency of references to sensor data decreases for an older rating index. Based on this finding, the sensor data deletion unit... 104In this embodiment, the storage priority of the sensor data decreases the longer the time elapsed between the measurement time and the current time, in order to make the sensor data more susceptible to deletion. This allows for the preferential deletion of sensor data for which the probability of user reference is low.

[0034] An example of a formula for determining the storage priority of the sensor data, reflecting the above finding, is shown as expression (2). P r ( x ) = f ( d e n ( x ) , d i f ( x ) , E l a ( x ) ) = d i f ( x ) E l a ( x ) × ( d e n ( x ) − 1 ) <?page 6=""?>

[0035] According to expression (2), the storage priority of the sensor data acquired at measurement time x is proportional to the degree of change in the anomaly level, inversely proportional to the elapsed time, and inversely proportional to the storage density. Regarding the storage density, it is set such that a number of data points stored in ±12 hours exceeding 1 can render the priority infinite.

[0036] The rating index storage unit 105 stores the time series anomaly degree data calculated by the valuation index calculation unit. 102 calculated in a specific memory area.

[0037] The user interface unit 106 The user interface unit displays the rating index and sensor data to the user by showing the rating index and sensor data on the screen or similar device. 106 provides resources to support the data management unit 1to instruct the display of the sensor data that formed the basis for calculating a specific rating index.

[0038] Fig. 3 and Fig. 4 are each an example of a user interface unit 106 displayed screen content. Fig. Figure 3 is a screen display showing a curve with the time series anomaly degree data plotted on it. The vertical axis of the curve represents the anomaly degree, and the horizontal axis represents time. When the user points to any point on the curve, the user interface unit identifies 106 The anomaly degree data (or the time) corresponding to the point being pointed to. The user interface unit. 106 records to the sensor data storage unit 103The system references and attempts to obtain the sensor data that corresponds to the identified anomaly level data (or time). If the corresponding sensor data exists, the user interface unit obtains it. 106 the sensor data and displays it to the in Fig. 4. Displayed screen content. Fig. Figure 4 displays a screen containing a curve with plotted time-series sensor data. The vertical axis of the curve represents the sensor data, and the horizontal axis represents time. If the corresponding sensor data does not exist, the user interface unit attempts to generate it. 106 iteratively, the system attempts to obtain the sensor data corresponding to other anomaly-degree data (or time) in the vicinity of the point being pointed at, until it succeeds in obtaining the sensor data. The user interface unit displays the obtained sensor data in the screen content of Fig. 4 dar.

[0039] Using the flowchart from Fig. 6. The operation of the data management device will be 1 described according to the embodiment.

[0040] S1: The sensor data acquisition unit 101 obtains the sensor data from one or more sensors 60 and collects the sensor data in the sensor data storage unit 103 .

[0041] At any time thereafter, the valuation index calculation unit calculates 102 using the sensor data storage unit 103 The collected time series sensor data determines the rating index, that is, the degree of anomaly.

[0042] S2: The sensor data erasure unit 104 monitors the total amount of sensor data stored in the sensor data storage unit 103 are stored. If the total amount exceeds the predetermined upper limit, processing proceeds to stepS3 Processing ends if the total quantity is equal to or less than the upper limit.

[0043] S3: The sensor data erasure unit 104 determined on the basis of at least one of the degrees of change in the valuation index, that is, the degree of anomaly determined by the valuation index calculation unit 102 The storage priority of the sensor data was calculated based on the storage density of the sensor data and the time that has passed since the sensor data was obtained.

[0044] S4: The sensor data erasure unit 104 deletes the sensor data with the lowest storage priority.

[0045] According to this embodiment, the storage priority of the sensor data is determined based on at least one of the following: the degree of change in the evaluation index (i.e., the degree of anomaly), the storage density of the sensor data, and the time elapsed since the sensor data was acquired. If the total amount of sensor data exceeds the predetermined upper limit, necessitating deletion, the sensor data with a low storage priority is deleted. This prevents unnecessary deletion of sensor data, or in other words, unnecessary thinning. Furthermore, it ensures that the size of the sensor data remains equal to or less than the predetermined upper limit.

[0046] According to this embodiment, the sensor data to be stored is determined solely based on the acquired sensor data and the degree of anomaly, thus eliminating the need for prior knowledge of anomalies and system failures. Furthermore, the feasibility of data erasure is assessed after a certain amount of sensor data has been stored, allowing statistics such as the variance of the anomaly degree to be used to determine this feasibility.

[0047] Although one embodiment of the invention has been described above, the invention is not limited to the example of the embodiment described above, but can be implemented in various ways with suitable modifications. For example, the sensor data erasure unit 104In the embodiment described above, a provision regarding the feasibility of deleting all sensor data stored in the sensor data storage unit 103 are stored. However, it is possible, for example, to implement a system where sensor data is excluded from the objects for determination during a predefined exception period (that is, so that it is not deleted). For example, the user can define an exception period for which it is desirable to permanently store the sensor data without thinning it out, such as the date and time of a change in the external environment. When a predefined event is detected, the data management unit can 1 to define a predetermined period, linked to the time of occurrence of the event, as the exceptional period.

[0048] In the embodiment described above, an example was shown in which the storage density and the storage density are designed to be inversely proportional to each other so that the sensor data erasure unit does not excessively thin out the sensor data. However, the invention is not limited to this example, and, for example, further conditions such as providing a lower limit for the number or amount of sensor data to be stored in a given period can be added to it.

[0049] Although the embodiment described above shows an example in which the user interface unit 106 The invention is not limited to the example of displaying the curves of the rating index and the sensor data. The user interface unit can display the rating index and the sensor data in any format, such as numerical values, a table, or the like.

[0050] Although the embodiment described above mainly represents an example in which the rating index calculation unit 102 The invention is not limited to this example, where the degree of anomaly is calculated as a rating index. The rating index calculation unit 102 The sensor data erasure unit can calculate any desired rating index using the sensor data. 104 Based on any rating index, it can determine whether or not the sensor data should be deleted. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] WO 2016 / 199210 [0004, 0005] JP 2017173321 [0004, 0005]

Claims

[1] Data management device that can calculate an evaluation index based on sensor data, wherein the data management device a rating index calculation unit that calculates the rating index using sensor data obtained from one or more sensors installed in an industrial machine; a sensor data storage unit that stores the sensor data; and a sensor data erasure unit that erases the sensor data from the sensor data storage unit when the total amount of sensor data in the sensor data storage unit exceeds an upper limit, includes, wherein the sensor data erasure unit determines a storage priority of the sensor data based on a degree of change in the rating index and erases the sensor data according to the storage priority. [2] Data management device according to claim 1, wherein the sensor data erasure unit uses a dispersion of the rating index or a time derivative of the rating index over a prescribed period as the degree of change of the rating index. [3] Data management device according to claim 1, wherein the sensor data erasure unit determines the storage priority of the sensor data based on the degree of change of the rating index and a storage density of the sensor data. [4] Data management device according to claim 1, wherein the sensor data erasure unit determines the storage priority of the sensor data based on the degree of change of the rating index and the time that has elapsed since the sensor data was obtained. [5] Data management device according to claim 1, further comprising: a user interface unit that displays time series data of the rating index, receives a selection of the rating index at a specific time, and displays time series data of the sensor data that formed a basis for calculating the selected rating index, wherein The user interface unit then displays other sensor data obtained at a time closest to the time the deleted sensor data was obtained, once the sensor data that formed the basis for calculating the selected rating index has been deleted. [6] Data management device according to claim 1, wherein the rating index is an anomaly level. [7] Data management program that causes a computer to perform a process of calculating a rating index based on sensor data and that causes the computer to execute a first step of calculating the rating index using the sensor data obtained from one or more sensors installed in an industrial machine; a second step of storing the sensor data in a sensor data storage unit; and a third step of deleting the sensor data from the sensor data storage unit if the total amount of sensor data in the sensor data storage unit exceeds an upper limit, brings, whereby in the third step a storage priority of the sensor data is determined based on a degree of change of the rating index and the sensor data are deleted according to the storage priority. [8] Data management procedures for calculating an evaluation index based on sensor data and for executing a first step of calculating the rating index using the sensor data obtained from one or more sensors installed in an industrial machine; a second step of storing the sensor data in a sensor data storage unit; and a third step of deleting the sensor data from the sensor data storage unit if the total amount of sensor data in the sensor data storage unit exceeds an upper limit, In the third step, a storage priority of the sensor data of the rating index is determined based on a degree of change, and the sensor data is deleted according to the storage priority.

Images