Information processing method and information processing device

Abstract

The present invention addresses the problem of providing an information processing method and an information processing device that enable more accurate alignment. This information processing method is characterized by including: an acquisition step for acquiring target data, which is time-series data including sampling values for each sampling timing, and standard time-series data, which is a reference for determining whether or not the target data is normal; a generation step in which, for each sampling value of the standard time-series data, a sampling timing corresponding to the sampling value is determined as a relative sampling timing that is a sampling timing measured from a predetermined starting point, and distribution data indicating a distribution range of the relative sampling timing is generated; and a determination step for determining a curved path based on the object data and the standard time series data so as to be included in a distribution range of the relative sampling timing determined based on the distribution data.

Description

Technical Field

[0001] The present invention relates to an information processing method and an information processing apparatus. Background Art

[0002] Conventionally, in production process management, in order to determine an abnormality of a production apparatus, a technique is known in which time series data of measured values such as temperature or pressure measured in the production apparatus is aligned with time series data in a normal state. For example, a technique is known in which, at the time of alignment, a tolerance range of a time deviation of a sampling timing is flexibly set for each sampled value.

[0003] [Prior Art Documents]

[0004] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent No. 5492622 Summary of the Invention

[0006] [Problems to be Solved by the Invention]

[0007] However, in the above-described prior art, in a process including a plurality of steps, in an alignment method using an elapsed time from a start time point of the process, there is a tendency that a wider distribution is adopted as the step is later in the process. Therefore, in a case where it is desired to detect a deviation of a required time of a step after the process as an abnormality, there is a case where, even when the required time of the step deviates from normal data as time series data in a normal state, it is determined to be within a range of normal data. In this case, there is a risk that the deviation of the required time of the step cannot be detected as an abnormality.

[0008] The present application is to solve such a problem, and an object thereof is to enable more accurate alignment.

[0009] [Means for Solving the Problems]

[0010] The information processing method of the present application is characterized by including: an acquisition step of acquiring object data that is time series data including sampled values at each sampling timing, and standard time series data that is a reference for determining whether the object data is normal; a generation step of, for each sampled value of the standard time series data, determining a sampling timing corresponding to the sampled value as a relative sampling timing that is a sampling timing measured from a specified start point, and generating distribution data indicating a distribution range of the relative sampling timing; and a determination step of determining a bending path based on the object data and the standard time series data so as to be included in the distribution range of the relative sampling timing determined based on the distribution data.

[0011] In the information processing method, the object data is compared with all the sampled values of the standard time series data, and the bending path that minimizes the distance between the object data and the standard time series data is determined as the optimal bending path.

[0012] In the information processing method, the bending path is determined by restricting the path of the bending path within the distribution range of the relative sampling timing determined based on the distribution data.

[0013] In the information processing method, a calculation process is further included, and the calculation process calculates the distance between the object data and the standard time series data by accumulating the values on the path of the bending path.

[0014] In the information processing method, a determination process is further included, and when the distance calculated by the calculation process is equal to or greater than a specified threshold value, it is determined that the object data is abnormal.

[0015] In the information processing method, the distribution data is the range of deviation of the value represented by the difference between the relative sampling timing and the sampling timing of the object data.

[0016] [Advantages of the Invention]

[0017] By the information processing method, object data as time series data including sampled values at each sampling timing and standard time series data as a reference when determining whether the object data is normal are obtained. For each sampled value of the standard time series data, the sampling timing corresponding to the sampled value is determined as the relative sampling timing as the sampling timing measured from a specified starting point, distribution data representing the distribution range of the relative sampling timing is generated, and the bending path based on the object data and the standard time series data is determined so as to be included in the distribution range of the relative sampling timing determined based on the distribution data. Therefore, more accurate alignment can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a conceptual diagram showing an outline of an embodiment.

[0019] Figure 2 It is a diagram showing a structural example of the information processing system of the embodiment.

[0020] Figure 3 It is a diagram showing an example of the data set information storage unit of the embodiment.

[0021] Figure 4 It is a diagram showing an example of the threshold information storage unit of the embodiment.

[0022] Figure 5 It is a conceptual diagram showing the generation process of the embodiment.

[0023] Figure 6 This is a flowchart showing an example of the process of a decision-making process executed by an information processing device according to an embodiment.

[0024] Figure 7 This is a flowchart showing an example of the process of a determination process executed by an information processing device according to an embodiment.

[0025] [Description of symbols]

[0026] N: Network

[0027] 1: Information processing system

[0028] 10: Data server

[0029] 100: Information processing device

[0030] 110: Communication unit

[0031] 120: Storage unit

[0032] 121: Dataset storage unit

[0033] 122: Threshold information storage unit

[0034] 130: Control unit

[0035] 131: Acquisition unit

[0036] 132: Generation unit

[0037] 133: Calculation unit

[0038] 134: Decision unit

[0039] 135: Judgment unit Detailed implementation manners

[0040] Next, the embodiments will be described with reference to the accompanying drawings. In addition, in the following description, the same reference numerals are assigned to the common constituent elements in each embodiment, and repeated descriptions are omitted.

[0041] [Examples]

[0042] [1. Outline of the embodiment]

[0043] At the site of production process management, in order to ensure the stable operation of production equipment, it is necessary to detect anomalies in production equipment as early as possible. In the past, in production process management, in order to determine anomalies in production equipment, there has been a technique of aligning time series data of measured values such as temperature or pressure measured in production equipment (time series data of a monitoring object for monitoring anomalies) with time series data in normal conditions. For example, there is a known technique of flexibly setting the allowable range of time deviation of sampling timing for each sampled value during alignment. In the alignment in this case, the dynamic time warping method (DTW) is used.

[0044] However, in the prior art, in a process including multiple processes, there is a risk of not being able to detect anomalies in measured values measured in production equipment. Generally, the operation data of a factory sometimes includes multiple processes. The data of the process in this case, for example, is data of a batch process in a factory, etc., and is various sensor data in the factory.

[0045] For example, the process includes: process 1 of putting in raw materials, process 2 of preprocessing the raw materials, process 3 of reacting the preprocessed raw materials, process 4 of ending the process, process 5 of taking out the product from the machine, and process 6 of shutting down the factory. Thus, in the present application, the process as an object is assumed to be a process with a clear start and end.

[0046] In this case, sometimes the time required until the end of the process changes due to the overall process or each process. For example, in normal data, there are also long data, short data, etc. in the time axis direction, and sometimes jitter occurs in the time axis direction. Therefore, synchronization processing in the time axis direction is required.

[0047] However, in the alignment method using the elapsed time from the start time point of the process, the later the process, the more likely it is to have a wider distribution. The reason for this tendency is considered to be that in the processes located later in the process, the distribution of the elapsed time from the start time point of the process includes not only the deviation of the required time of the process but also the deviations of all the processes that have elapsed so far. Therefore, in the case of wanting to detect the deviation of the required time of a process located later as an anomaly, even when the required time of the process deviates from the normal data, it is sometimes determined to be within the range of normal data in the distribution including the deviations of the required times of the previous processes that have elapsed so far, and thus sometimes the deviation of the required time of the process cannot be detected as an anomaly.

[0048] Here, use Figure 1 to illustrate the deviation of the required time of the process. Figure 1 is a conceptual diagram showing an overview of the embodiment. In Figure 1In the example, FIG. FI1 shows time series data with the horizontal axis representing time and the vertical axis representing sensor values (an example of sampled values). Additionally, in FIG. FI1, it is a graph formed by overlapping multiple time series data. For example, in FIG. FI1, the dashed line represents normal data, and the solid line represents the object data OD1. The normal data mentioned here is, for example, time series data of normal sensor values obtained from a factory.

[0049] The process in this case includes: process ST1, process ST2, process ST3, process ST4, and process ST5. Here, as Figure 1 shown, it can be seen that the deviation FL1 in process ST5 is larger than that in other processes. Thus, in the required time for each of the processes ST1 to ST5, when wanting to determine normal or abnormal, or when wanting to determine normal or abnormal in the required time for all processes, there is a problem that the deviation of the required time becomes larger in subsequent processes (in Figure 1 the example, for example, corresponding to process ST5).

[0050] In such time series data, if the existing technology is applied, even when the deviation of the sampling timing of the object data from the time axis direction of the normal data deviates, there is a possibility that the sampling timing of the standard time series data synchronizes with the sampling timing of the object data when using DTW for alignment. In addition, the sampling timing represents a real moment, or the elapsed time since the start of the process, or an index of time, etc.

[0051] Therefore, in the present application, to address the above problem, a relative sampling timing measured from a specified starting point is introduced. For example, the present application provides a solution for generating distribution data based on the relative sampling timing. For example, in the present application, a point selected from any data point in the sampling timing of the time series data is used as the starting point of the relative sampling timing. Moreover, the present application determines the range of the difference between the current sampling timing and the starting point as the range that can be synchronized. And the present application performs alignment based on the range that can be synchronized.

[0052] Hereinafter, an example of a sampling timing where the starting point of the relative sampling timing is a value specified to be earlier than the current sampling timing will be described.

[0053] 〔2. Structure of the Information Processing System〕

[0054] Use Figure 2 to illustrate an example of the structure of the information processing system 1. Figure 2 is a diagram showing an example of the structure of the information processing system 1 according to the embodiment. In Figure 2In the example, the information processing system 1 includes a data server 10 and an information processing device 100. The data server 10 and the information processing device 100 are communicably connected via the network N by wire or wirelessly. In addition, Figure 2 The information processing system 1 shown may also include multiple data servers 10 or multiple information processing devices 100.

[0055] The data server 10 is an information processing device that acquires various data and provides them to the information processing device 100, and is implemented, for example, by a server device or a cloud system, etc. For example, the data server 10 stores object data newly acquired from a factory. For example, the object data is time-series data including sampling values at each sampling timing. In this case, the data server 10 may also store the information related to the process corresponding to the sampling timing of the object data in association with the object data.

[0056] In addition, the data server 10 stores standard time-series data. Here, the standard time-series data is data generated based on normal data, which is time-series data representing normal sensor values acquired from the factory. For example, the standard time-series data is data used as a reference when determining whether the object data newly acquired from the factory is normal.

[0057] The standard time-series data is data generated according to the average value of each of multiple normal data. If more specific examples are given, the value of the sampling timing (the value on the time axis) of the standard time-series data is the average value of the value of the sampling timing of the first normal data and the value of the sampling timing of the second normal data. In addition, the sensor value of the standard time-series data is the average value of the sensor value of the first normal data and the sensor value of the second normal data.

[0058] In addition, in the above example, the value of the sampling timing of the standard time-series data may not be limited to the example of the average value of the value of the sampling timing of the first normal data and the value of the sampling timing of the second normal data. For example, instead of the average value, it may be a weighted average or the like.

[0059] The information processing device 100 is an information processing device that can communicate with various devices via the network N, and is implemented, for example, by a server device or a cloud system, etc. For example, the information processing device 100 is connected to various other devices via the network N in a communicable manner.

[0060] For example, the information processing apparatus 100 acquires object data and standard time series data from the data server 10. Subsequently, the information processing apparatus 100 generates distribution data representing the distribution range of the relative sampling timing, which is the sampling timing measured from a specified starting point. Then, the information processing apparatus 100 determines a bending path based on the object data and the standard time series data so that it is included in the distribution range of the relative sampling timing determined based on the generated distribution data.

[0061] That is, the information processing apparatus 100 determines the range of the difference between the current sampling timing and the starting point of the relative sampling timing as the range that can be synchronized, and performs alignment based on the range that can be synchronized. As a result, the information processing apparatus 100 can achieve more accurate alignment.

[0062] [3. Structure of the Information Processing Apparatus]

[0063] Next, an example of the functional structure of the information processing apparatus 100 will be described. As Figure 2 shown, the information processing apparatus 100 of the embodiment includes: a communication unit 110, a storage unit 120, and a control unit 130.

[0064] (Regarding the Communication Unit 110)

[0065] The communication unit 110 is implemented, for example, by a Network Interface Card (NIC) or the like. Moreover, the communication unit 110 is connected to the network N by wire or wirelessly, and performs information transmission and reception with various other devices.

[0066] (Regarding the Storage Unit 120)

[0067] The storage unit 120 is implemented, for example, by semiconductor storage elements such as a Random Access Memory (RAM) and a Flash Memory, or storage devices such as a hard disk and an optical disk. For example, the storage unit 120 includes a data set storage unit 121 and a threshold information storage unit 122.

[0068] (Regarding the Data Set Storage Unit 121)

[0069] The data set storage unit 121 stores various data. Here, an example of the data set storage unit 121 of the embodiment is shown in Figure 3 . In the example shown in Figure 3 , the data set storage unit 121 has items such as "Data Set ID (Identifier)", "Standard Time Series Data ID", "Sampling Timing", "Sensor Value", "Object Data ID", "Sampling Timing", "Sensor Value", "Relative Distribution Data ID", "Relative Distribution Data", etc.

[0070] "Dataset ID" is an identifier that identifies a dataset as a set of data. "Standard time series data ID" is an identifier that identifies the standard time series data corresponding to the "Dataset ID". "Sampling timing" is information related to the sampling timing of the standard time series data corresponding to the "Standard time series data ID". For example, the sampling timing of the standard time series data is the time measured with the start point of the process as 0. "Sensor value" is information related to the sensor value of the standard time series data corresponding to the "Standard time series data ID".

[0071] "Object data ID" is an identifier that identifies the object data corresponding to the "Dataset ID". "Sampling timing" is information related to the sampling timing of the object data corresponding to the "Object data ID". For example, the sampling timing of the object data is the time measured with the start point of the process as 0. "Sensor value" is information related to the sensor value of the object data corresponding to the "Object data ID".

[0072] "Distribution data ID" is an identifier that identifies the distribution data corresponding to the "Dataset ID". "Distribution data" is information related to the distribution data corresponding to the "Distribution data ID". For example, the distribution data is the range of deviation of the value represented by the difference between the relative sampling timing and the sampling timing of the normal data.

[0073] For example, in Figure 3 , in "D1" identified by the dataset ID, the standard time series data ID is "SD1", the sampling timing is "ST1", and the sensor value is "SS1". In addition, the object data ID is "OD1", the sampling timing is "DT1", and the sensor value is "DS1". In addition, the distribution data ID is "AD1", and the distribution data is "ADD1".

[0074] In addition, in Figure 3 , in the example shown, the sampling timing etc. are represented by abstract symbols such as "ST1", but the sampling timing etc. can also be in the form of a file containing specific numerical values, or specific character strings, or various information representing the sampling timing etc.

[0075] (Regarding the threshold information storage unit 122)

[0076] The threshold information storage unit 122 stores information related to a prescribed threshold. Here, an example of the threshold information storage unit 122 of the embodiment is shown in Figure 4 . In the example shown in Figure 4 , the threshold information storage unit 122 has items such as "Threshold ID" and "Threshold information".

[0077] The "threshold ID" is an identifier for identifying threshold information. The "threshold information" is information related to the threshold corresponding to the "threshold ID". For example, the information related to the threshold represents a specified threshold. For example, the specified threshold is a value estimated from time series data at the time of an abnormality, and is an arbitrary value. In addition, the specified threshold may also be set by considering the distance between the time series data at the time of an abnormality and the standard time series data. In this case, the specified threshold is the lower limit value of the calculated distance.

[0078] For example, in Figure 4 , in "T1" identified by the threshold ID, the threshold information is "TH1". In addition, in the example shown in Figure 4 , the threshold information, etc. are represented by abstract symbols such as "TH1", but the threshold information, etc. may also be in the form of a file, etc. that includes specific numerical values, or specific character strings, or various information representing the threshold information, etc.

[0079] (Regarding the control unit 130)

[0080] The control unit 130 is a controller, and is implemented, for example, by using a central processing unit (CPU) or a micro processing unit (MPU) to execute with the RAM as the working area through various programs stored in the storage device inside the information processing device 100. In addition, the control unit 130 is a controller, and is implemented, for example, by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA).

[0081] As shown in Figure 2 , the control unit 130 has an acquisition unit 131, a generation unit 132, a calculation unit 133, a determination unit 134, and a judgment unit 135, and realizes or executes the functions or operations of the information processing described below. In addition, the internal structure of the control unit 130 is not limited to the structure shown in Figure 2 , and may also be other structures as long as it is a structure for performing the information processing described later. In addition, the connection relationship of each processing unit included in the control unit 130 is not limited to the connection relationship shown in Figure 2 , and may also be other connection relationships.

[0082] (Regarding the acquisition unit 131)

[0083] The acquisition unit 131 acquires various information. Specifically, the acquisition unit 131 acquires object data that is time-series data including sampled values at each sampling timing and standard time-series data that is a reference when determining whether the time-series data is normal.

[0084] For example, the acquisition unit 131 acquires object data and standard time-series data from the data server 10. Then, the acquisition unit 131 stores the acquired object data and standard time-series data in the data set storage unit 121.

[0085] (Regarding the generation unit 132)

[0086] The generation unit 132 generates various information. Specifically, for each sampled value of the standard time-series data, the generation unit 132 determines the sampling timing corresponding to the sampled value as a relative sampling timing that is a sampling timing measured from a specified starting point. In this case, the generation unit 132 determines the starting point of the relative sampling timing for each sampling timing of the standard time-series data.

[0087] For example, the generation unit 132 determines the sampling timing that is a specified value earlier than the current sampling timing as the starting point of the relative sampling timing.

[0088] Subsequently, the generation unit 132 generates distribution data representing the distribution range of the determined relative sampling timing. Then, the generation unit 132 stores the generated distribution data in the data set storage unit 121.

[0089] Subsequently, the generation unit 132 determines the range corresponding to the sampling timing based on the distribution data. For example, the generation unit 132 determines the distribution range of the relative sampling timing represented by the distribution data as the range corresponding to the sampling timing.

[0090] Here, use Figure 5 To explain the generation process of the distribution data. Figure 5 is a conceptual diagram showing the generation process of the embodiment. In Figure 5 example, Table TA1 shows the sampling timing of the standard time-series data and the sampling timing of normal data 1 to normal data N (for example, N is an arbitrary integer) that is the basis of the standard time-series data.

[0091] Here, in Table TA1, when the sampling timing of the standard time-series data is "1", the sampling timing of normal data 1 is "1", the sampling timing of normal data 2 is "1", and the sampling timing of normal data N is "1".

[0092] In addition, in Table TA1, when the sampling timing of the standard time series data is "20", the sampling timing of normal data 1 is "17", the sampling timing of normal data 2 is "22", and the sampling timing of normal data N is "19".

[0093] In addition, in Table TA1, when the sampling timing of the standard time series data is "35", the sampling timing of normal data 1 is "28", the sampling timing of normal data 2 is "33", and the sampling timing of normal data N is "38".

[0094] Here, it is assumed that the starting point of the relative sampling timing is the sampling timing "20" of the standard time series data. In this case, the relative sampling timing is represented by the difference between the sampling timing as the starting point and the current sampling timing. In Figure 5 the example, the relative sampling timing is represented by the difference from the sampling timing "35" to the sampling timing "20" of the standard time series data.

[0095] Subsequently, distribution data of the relative sampling timing corresponding to the sampling timing of the standard time series data is generated. First, the difference in the sampling timing of each normal data corresponding to the relative sampling timing is calculated. In Figure 5 the example, the difference between the sampling timing "17" of normal data 1 corresponding to the sampling timing of the starting point and the sampling timing "28" of normal data 1 corresponding to the current sampling timing is "+11".

[0096] In addition, the difference between the sampling timing "22" of normal data 2 corresponding to the sampling timing of the starting point and the sampling timing "33" of normal data 2 corresponding to the current sampling timing is "+11".

[0097] In addition, the difference between the sampling timing "19" of normal data N corresponding to the sampling timing of the starting point and the sampling timing "38" of normal data N corresponding to the current sampling timing is "+19".

[0098] Next, the generation unit 132 generates distribution data based on the differences in the sampling timing of each normal data. In Figure 5 the example, the generation unit 132 generates the distribution data in normal data 1 to normal data N as "+11 to +19". In this way, the generation unit 132 generates the distribution data "+11 to +19" of the relative sampling timing corresponding to the sampling timing "35" of the standard time series data. Then, the generation unit 132 stores the generated distribution data "+11 to +19" in the data set storage unit 121.

[0099] In addition, the distribution data may, for example, also include the minimum value of the sampling timing and the maximum value of the sampling timing. In addition, the distribution data may also represent the standard deviation of the sampling timing.

[0100] (Regarding the calculation unit 133)

[0101] The calculation unit 133 calculates a distance matrix based on the object data and the standard time series data. In this case, in the distance matrix, the rows correspond to the standard time series data and the columns correspond to the object data. For example, as an example of the absolute difference between the sensor value of the object data and the sensor value of the standard time series data, the calculation unit 133 calculates the absolute difference in all combinations of all sensor values of the object data and all sensor values of the standard time series data. In this way, the calculation unit 133 calculates the distance matrix based on the object data and the standard time series data. Then, the calculation unit 133 stores the calculated distance matrix in the storage unit 120.

[0102] In addition, the calculation unit 133 calculates the distance between the object data and the standard time series data by accumulating the values on the path of the bending path. For example, if an example of a pre-determined bending path is used for explanation, the calculation unit 133 uses the values on the path of the bending path to calculate the distance between the object data and the standard time series data. For example, the calculation unit 133 refers to the distance matrix stored in the storage unit 120, accumulates the absolute differences of the respective matrix elements forming the bending path, and calculates the distance between the standard time series data and the object data. The matrix elements mentioned here correspond to the matrix elements of the distance matrix based on the object data and the standard time series data. Then, the calculation unit 133 stores the calculated distance in the storage unit 120.

[0103] (Regarding the determination unit 134)

[0104] The determination unit 134 determines a bending path based on the object data and the standard time series data so as to be included in the distribution range of the relative sampling timing determined based on the distribution data.

[0105] For example, the determination unit 134 compares all the sampling values of the object data and the standard time series data with each other, and determines the bending path that minimizes the distance between the object data and the standard time series data. In this case, the determination unit 134 determines the bending path by restricting the path of the bending path within the distribution range of the relative sampling timing determined based on the distribution data stored in the data set storage unit 121. If the example of Figure 5 is used for explanation, the determination unit 134 determines the bending path by restricting the path of the bending path within Figure 5 the range of the distribution data “+11 to +19” generated by the generation unit 132 in

[0106] If more specific examples are enumerated, for example, the determination unit 134 refers to the distance matrix calculated by the calculation unit 133 and stored in the storage unit 120, and determines to exclude matrix elements included in the bending path that are outside the distribution range relative to the sampling timing from the paths of the bending path.

[0107] For example, it is assumed that one matrix element included in the bending path is outside the distribution range relative to the sampling timing. In this case, the determination unit 134 changes one matrix element included in the bending path to a matrix element in the same column and different rows, and is a matrix element included in the bending path within the distribution range relative to the sampling timing. In this way, the determination unit 134 determines the bending path restricted within the distribution range relative to the sampling timing.

[0108] (Regarding the determination unit 135)

[0109] The determination unit 135 determines that the target data is abnormal when the distance calculated by the calculation unit 133 is equal to or greater than a specified threshold value. For example, the determination unit 135 determines that the target data is abnormal when the distance calculated by the calculation unit 133 and stored in the storage unit 120 is equal to or greater than the specified threshold value stored in the threshold information storage unit 122. On the other hand, the determination unit 135 determines that the target data is normal when the distance stored in the storage unit 120 is less than the specified threshold value stored in the threshold information storage unit 122.

[0110] [4. Processing flow (1)]

[0111] Next, use Figure 6 to describe the flow of the determination process executed by the information processing apparatus 100 of the embodiment. Figure 6 is a flowchart showing an example of the flow of the determination process executed by the information processing apparatus 100 of the embodiment.

[0112] For example, it is assumed in advance that standard time series data and target data are obtained from the data server 10. In addition, it is assumed that alignment has been performed at all points of the standard time series data and the normal data.

[0113] In this case, as Figure 6 shown, the information processing apparatus 100 substitutes 1 for the sampling timing t (step S101). Subsequently, the acquisition unit 131 acquires information related to the sampling timing of the target data from the data set storage unit 121 (step S102).

[0114] Then, the generation unit 132 selects the starting point of the relative sampling timing among the sampling timings of the target data (step S103). For example, the generation unit 132 determines a sampling timing that is a specified value earlier than the current sampling timing as the starting point of the relative sampling timing.

[0115] Subsequently, the acquisition unit 131 acquires the sampling timing corresponding to the start point of the relative sampling timing in the standard time series data (step S104). For example, the acquisition unit 131 acquires the sampling timing of the standard time series data corresponding to the relative sampling timing of the target data.

[0116] Then, the generation unit 132 generates distribution data based on the relative sampling timing (step S105). For example, the generation unit 132 generates distribution data representing the distribution range of the relative sampling timing.

[0117] Subsequently, the generation unit 132 determines the range corresponding to the sampling timing based on the distribution data (step S106). For example, the generation unit 132 determines the distribution range of the relative sampling timing represented by the distribution data as the range corresponding to the sampling timing.

[0118] Then, the determination unit 134 determines whether to associate the target data based on the corresponding range (step S107). For example, the determination unit 134 compares all the sampling values of the target data with the standard time series data, and determines the warping path with the smallest distance between the target data and the standard time series data as the optimal warping path. In this case, the determination unit 134 determines the warping path by restricting the path of the warping path to the distribution range of the relative sampling timing determined based on the distribution data stored in the data set storage unit 121. That is, the determination unit 134 performs DTW within the distribution range to search for the optimal point.

[0119] If a more specific example is cited, for example, the determination unit 134 refers to the distance matrix stored in the storage unit 120 and determines to exclude the matrix elements included in the warping paths located outside the distribution range of the relative sampling timing from the paths of the warping paths.

[0120] For example, assume that one of the matrix elements included in the warping path is located outside the distribution range of the relative sampling timing. In this case, the determination unit 134 changes one of the matrix elements included in the warping path to a matrix element in the same column and different row, and is a matrix element included in the warping path located within the distribution range of the relative sampling timing. In this way, the determination unit 134 determines the warping path restricted to the distribution range of the relative sampling timing.

[0121] Subsequently, the information processing apparatus 100 substitutes t + 1 for the sampling timing t (step S108). Then, the information processing apparatus 100 determines whether t is greater than a specified value (step S109). For example, the specified value is the value representing the data length of the target data.

[0122] Specifically, when the information processing apparatus 100 determines that t is less than a specified value (step S109; No), it returns to step S102. On the other hand, when the information processing apparatus 100 determines that t is greater than the specified value (step S109; Yes), the information processing ends.

[0123] In addition, in the above example, an example in which the information processing apparatus 100 substitutes 1 for the sampling timing t in step S101 is described, but it is not limited thereto, and it may be executed by various structures (such as the acquisition unit 131, etc.) provided in the information processing apparatus 100. Further, an example in which the information processing apparatus 100 substitutes t + 1 for the sampling timing t in step S108 is described, but it is not limited thereto, and it may be executed by various structures provided in the information processing apparatus 100. Additionally, an example in which the information processing apparatus 100 determines whether t is greater than a specified value in step S109 is described, but it is not limited thereto, and it may be executed by various structures provided in the information processing apparatus 100.

[0124] Furthermore, in the above example, an example in which the information processing that repeats t to a specified value is described, but it is not limited thereto. For example, the above information processing may be performed only on the sampling timing of a part of the target data.

[0125] 〔5. Processing flow (2)〕

[0126] Next, Figure 7 the flow of the determination process executed by the information processing apparatus 100 of the embodiment will be described. Figure 7 FIG. is a flowchart showing an example of the flow of the determination process executed by the information processing apparatus 100 of the embodiment.

[0127] As Figure 7 shown, the calculation unit 133 calculates the distance between the target data and the standard time series data (step S201). For example, the calculation unit 133 accumulates the absolute differences on the path of the bending path between the target data and the standard time series data, and calculates the distance between the target data and the standard time series data.

[0128] Subsequently, the determination unit 135 determines whether the distance calculated by the calculation unit 133 is equal to or greater than a specified threshold value (step S202). Specifically, when the distance is equal to or greater than the specified threshold value (step S202; Yes), the determination unit 135 determines that the target data is abnormal (step S203).

[0129] On the other hand, when the distance is less than the specified threshold value (step S202; No), the determination unit 135 determines that the target data is normal (step S204).

[0130] 〔6. Effects〕

[0131] As described above, the information processing apparatus 100 of the embodiment acquires object data and standard time series data. Subsequently, for each sampled value of the standard time series data, the information processing apparatus 100 determines the sampling timing corresponding to the sampled value as the relative sampling timing. Then, the information processing apparatus 100 generates distribution data representing the distribution range of the relative sampling timing. Subsequently, the information processing apparatus 100 determines a bending path based on the object data and the standard time series data so that it is included in the distribution range of the relative sampling timing determined based on the distribution data.

[0132] For example, by using the relative sampling timing, the information processing apparatus 100 can correctly evaluate the deviation of the required time for each process regardless of the position on the time axis. In addition, when alignment is performed using DTW, the generated distribution data becomes a constraint condition. Then, when aligning the object data with the standard time series data, the appropriate points based on the distribution data among the points of the object data and the points of the standard time series data are associated with each other. Thereby, the information processing apparatus 100 can perform appropriate synchronization processing conforming to the distribution of normal data. In this way, the information processing apparatus 100 can achieve more accurate alignment.

[0133] In addition, since the information processing apparatus 100 determines the sampling timing that is a predetermined value ahead of the current sampling timing as the starting point of the relative sampling timing and determines the relative sampling timing, it is possible to avoid the influence of the deviation of the required time for the process before the process including sampling.

[0134] 〔7. Modification Example〕

[0135] The information processing apparatus 100 can be implemented in various different forms in addition to the above-described embodiment. Therefore, another embodiment of the information processing apparatus 100 will be described below.

[0136] 〔7-1. Sampled Value〕

[0137] In the above-described embodiment, an example in which the sampled value is a sensor value has been described, but it is not limited thereto. For example, the sampled value may be any value representing a measured value such as temperature or pressure measured in a production facility.

[0138] 〔7-2. Starting Point of Relative Sampling Timing〕

[0139] In the above-described embodiment, an example in which the starting point of the relative sampling timing is the sampling timing that is a predetermined value ahead of the current sampling timing has been described, but it is not limited thereto.

[0140] For example, the starting point of the relative sampling timing can also be the sampling timing corresponding to the time obtained from the actual operation process data. For example, the starting point can also be the sampling timing corresponding to the time recorded as a signal of the process operation. If more specific examples are listed, the information processing device 100 can determine the first sampling timing having the same value as the sampling value in the sampling timing as the starting point of the relative sampling timing in the identification information of the process obtained in the process.

[0141] In addition, the starting point of the relative sampling timing can also be estimated by machine learning. For example, the information processing device 100 determines a point where the trajectory of the data in the time series data changes significantly based on the existing technologies related to machine learning. Moreover, the information processing device 100 can also determine the point closest to the sampling timing among the determined points as the starting point of the relative sampling timing. In addition, the starting point of the relative sampling timing is preferably set to a point close to the starting point of the process including the sampling.

[0142] 〔7-3. Information processing device〕

[0143] In addition, the information processing device 100 can also be an information processing device for the monitored object data. For example, the information processing device 100 can also determine whether the time series data including sampling values such as temperature or pressure obtained from the data server 10 is normal. For example, the determination unit 135 determines that the object data is abnormal when the distance calculated by the calculation unit 133 is equal to or greater than the specified threshold value stored in the threshold information storage unit 122. On the other hand, the determination unit 135 determines that the object data is normal when the distance calculated by the calculation unit 133 is less than the specified threshold value stored in the threshold information storage unit 122.

[0144] In addition, the determination process performed by the information processing device 100 can also determine whether the object data is normal when the degree of consistency between the aligned object data and the aligned normal data is less than the specified threshold value.

[0145] For example, the calculation unit 133 calculates the degree of consistency between the aligned object data and the aligned normal data. In addition, for the calculation process, various existing technologies for calculating the degree of consistency between data can be adopted.

[0146] Moreover, the determination unit 135 can also determine that the object data is abnormal when the degree of consistency calculated by the calculation unit 133 is less than the specified threshold value. On the other hand, the determination unit 135 can also determine that the object data is normal when the degree of consistency calculated by the calculation unit 133 is equal to or greater than the specified threshold value.

[0147] 〔7-4. Relative sampling timing〕

[0148] In addition, the information processing apparatus 100 may also synthesize the sampling timing and the relative sampling timing, such as the product set, etc., as a new relative sampling timing.

[0149] [8. Others]

[0150] In addition, in each of the processes described in the above embodiments and modification examples, all or part of the processes described as being automatically performed may be manually performed, or all or part of the processes described as being manually performed may be automatically performed by a known method. In addition, regarding the information including the processing flow, specific names, various data, or parameters shown in the above document or drawings, unless otherwise specified, it may be arbitrarily changed. For example, the various information shown in each figure is not limited to the information shown in the figure.

[0151] In addition, each constituent component of each device shown in the figure is a functionally conceptual constituent component, and it is not necessarily physically configured as shown in the figure. That is, the specific form of dispersion and integration of each device is not limited to the form shown in the figure, and all or part of it may be functionally or physically dispersed and integrated in any unit according to various loads or usage conditions, etc.

[0152] In addition, the above embodiments and modification examples can be appropriately combined within a range where the processing contents do not conflict.

[0153] In addition, the so-called "section (part, module, unit)" can be replaced with "component" or "circuit", etc. For example, the determination section can be replaced with a determination component or a determination circuit.

[0154] As described above, several embodiments of the present application have been described in detail based on the drawings, but these are examples, and the present invention can be implemented in other forms in which various deformations and improvements have been made based on the knowledge of those skilled in the art, starting from the form described in the disclosure section of the invention.

Claims

1. An information processing method, characterized in that, Comprising: An acquisition step of acquiring object data which is time series data including sampling values at each sampling timing, and reference time series data which is a reference when determining whether the object data is normal; A generation step of, for each of the sampling values of the reference time series data, determining the sampling timing corresponding to the sampling value as a relative sampling timing which is a sampling timing measured from a specified starting point, and generating distribution data representing the distribution range of the relative sampling timing; And A determination step of determining a bending path based on the object data and the reference time series data so that it is included in the distribution range of the relative sampling timing determined based on the distribution data.

2. The information processing method according to claim 1, characterized in that The determination step Compares all of the sampling values of the object data and the reference time series data with each other, and determines the bending path that minimizes the distance between the object data and the reference time series data as the bending path.

3. The information processing method according to claim 2, characterized in that The determination step Determines the bending path by restricting the path of the bending path within the distribution range of the relative sampling timing determined based on the distribution data.

4. The information processing method according to claim 2, wherein It further comprises a calculation step, The calculation step calculates the distance between the object data and the reference time series data by accumulating the values on the path of the bending path.

5. The information processing method according to claim 4, wherein It further comprises a determination step, The determination step determines that the object data is abnormal when the distance calculated by the calculation step is equal to or greater than a specified threshold value.

6. The information processing method according to claim 1, characterized in that The distribution data is a range of deviation of values represented by the difference between the relative sampling timing and the sampling timing of normal data which is time series data in a normal state.

7. An information processing apparatus, characterized in that, Including: An acquisition unit that acquires object data which is time series data including sampling values at each sampling timing, and reference time series data which is a reference when determining whether the object data is normal; A generation unit that, for each of the sampling values of the reference time series data, determines the sampling timing corresponding to the sampling value as a relative sampling timing which is a sampling timing measured from a specified starting point, and generates distribution data representing the distribution range of the relative sampling timing; And A determination unit that determines a bending path based on the object data and the reference time series data so that it is included in the distribution range of the relative sampling timing determined based on the distribution data.

Images