Cause analysis device
The cause analysis device addresses the inefficiency in identifying multiple cause factors of abnormal states by using time-series data and interactive prompts, reducing man-hours and improving accuracy.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- FANUC LTD
- Filing Date
- 2025-01-14
- Publication Date
- 2026-07-23
AI Technical Summary
Existing methods require significant man-hours to identify the cause of abnormal states in industrial machine control devices due to multiple factors satisfying the occurrence condition, necessitating a more efficient technique for cause analysis.
A cause analysis device that acquires time-series data and abnormal state information, generating stepwise prompts for an interactive response device to identify the cause of the abnormal state, incorporating feedback for improved accuracy.
Reduces man-hours required for identifying causes of abnormal states by efficiently generating prompts and utilizing user feedback for enhanced accuracy.
Smart Images
Figure JP2025000892_23072026_PF_FP_ABST
Abstract
Description
Cause analysis device , ,
[0004] , ,
[0003] ,
[0006] ,
[0005] ,
[0001] An embodiment of the present invention relates to a technique for analyzing the cause of an abnormal state in a control device of an industrial machine.
[0002] Among the abnormal states of the control device, even if there is only one occurrence condition, there are cases where there are multiple factors that satisfy this condition. For example, an alarm indicating an abnormal state of excessive error during movement occurs when the deviation between the command position and the machine position exceeds the limit value. The factors include a small limit value setting, an excessive feed speed, insufficient adjustment of the position gain, an error in the detection unit, input of a servo off signal, an abnormal power supply voltage, etc., and there are multiple of them. Therefore, the investigation of the cause of the abnormal state is carried out for each of these factors, and it is necessary to investigate the operation history and input / output data of the control device retroactively from the time of occurrence of the abnormal state.
[0003] Japanese Unexamined Patent Application Publication No. 2021-9584
[0004] Since there are multiple factors that satisfy the occurrence condition of the abnormal state and it is necessary to investigate the data retroactively from the time of occurrence of the abnormal state, a lot of man-hours have been required to identify the cause. Therefore, a technique that can easily identify the cause is desired.
[0005] A cause analysis device according to an embodiment is a device for identifying the cause of an abnormal state in a control device. The cause analysis device acquires one or more time-series data in which an identifier for representing the specific abnormal state in the control device, or abnormal state information including information explaining the abnormal state, and information related to the control by the control device are linked to time data. Next, the cause analysis device generates a prompt for sequentially instructing each of the step-by-step procedures for identifying the cause of the abnormal state based on the abnormal state information and the time-series data, and transmits the prompt to the interactive response device. When the cause analysis device receives a response to the prompt from the interactive response device, it notifies the user of the content of the response.
[0006] This is a block diagram illustrating the functional configuration of the cause analysis device in the embodiment. This is a diagram illustrating the abnormal state information acquired by the cause analysis device in Operation Example 1. This is a diagram illustrating the time-series data acquired by the cause analysis device in Operation Example 1. This is a diagram illustrating the prompt generated by the cause analysis device and the response of the interactive response device in Operation Example 1. This is a diagram illustrating the prompt generated by the cause analysis device and the response of the interactive response device in Operation Example 2. This is a diagram illustrating the abnormal state information acquired by the cause analysis device in Operation Example 3. This is a diagram illustrating the time-series data acquired by the cause analysis device in Operation Example 3. This is a diagram illustrating the prompt generated by the cause analysis device and the response of the interactive response device in Operation Example 3.
[0007] The following describes embodiments of the present invention. The cause analysis device of this embodiment acquires abnormal state information and various time-series data from a control device such as a numerical control device or a robot control device, and uses these to input prompts in stages to an interactive response device (generating AI) to identify the cause of the abnormal state in the control device.
[0008] Figure 1 is a block diagram showing the functional configuration of the cause analysis device 1 in the embodiment. The cause analysis device 1 is an information processing device (computer) equipped with a control unit 10 (processor) and a storage unit 20 (memory), and is an auxiliary device for identifying the cause of an abnormal state that occurs in the control device 2. The cause analysis device 1 is also connected to the interactive response device 3 either directly or via a network.
[0009] The control unit 10 comprises a data acquisition unit 11, a prompt generation unit 12, a transmission unit 13, a reception unit 14, and a notification unit 15. These functional units operate when the control unit 10 reads and executes a cause analysis program stored in the storage unit 20.
[0010] The memory unit 20 stores various software, including the cause analysis program, as well as input / output data, data in progress, and so on.
[0011] The data acquisition unit 11 acquires one or more time-series data sets linked to time data, including an identifier for representing a unique abnormal state in the control device 2, or abnormal state information including information describing the abnormal state, and information related to control by the control device 2. The data acquisition unit 11 may acquire the abnormal state information and time-series data from the control device 2 in response to the occurrence of an abnormal state, but is not limited to this, and may acquire them directly or indirectly from the control device 2 at all times or periodically.
[0012] The abnormal state information may include manual information for the control device 2, inquiry information regarding the control device 2, source code information for the control device 2, the processing program in which the abnormal state occurred, the operation history information of the control device 2 performed up to the occurrence of the abnormal state, and countermeasures for each of the possible causes of the abnormal state. By adding this information to the prompt, a more appropriate response regarding the cause of the abnormal state and countermeasures can be expected.
[0013] The time-series data includes operating information of the control device 2, data represented by physical quantities detectable by the control device 2, signals input and output to the control device 2, and variables used in the processing program of the control device 2. Furthermore, the time-series data may also include data of normal conditions that can be compared with abnormal conditions, thereby clarifying the characteristics of the time-series data in abnormal conditions.
[0014] The prompt generation unit 12 generates prompts that sequentially instruct the interactive response device 3 to perform each of the stepwise steps for identifying the cause of the abnormal state based on the abnormal state information and time-series data. Specifically, in order to identify the cause of the abnormal state, the following stepwise steps are required: (1) Select time-series data related to the abnormal state based on the abnormal state information; (2) Present the characteristics of the changes in the time-series data; (3) Present the cause of the abnormal state based on the abnormal state information and the characteristics of the changes in the time-series data. The prompt generation unit 12 generates prompts that instruct the interactive response device 3 to perform at least one of these steps.
[0015] In this case, the prompt generation unit 12 may generate a prompt that instructs the interactive response device 3 to present a graph of time-series data that includes the portion from which the characteristics of change have been extracted.
[0016] Furthermore, the prompt generation unit 12 may instruct the interactive response device 3 to present a method for dealing with the identified cause of the abnormal state.
[0017] Furthermore, the prompt generation unit 12 may receive feedback from the user regarding the correctness of the notified response and generate a prompt that instructs the interactive response device 3 to present an alternative cause for the abnormal state that is different from the response.
[0018] The transmitting unit 13 transmits the prompt generated by the prompt generation unit 12 to the interactive response device 3. The receiving unit 14 receives the response to the prompt from the interactive response device 3. The notification unit 15 notifies the user of the content of the response received by the receiving unit 14 from the interactive response device 3.
[0019] Here, we show examples 1 to 3 of the operation in which the cause analysis device 1 generates prompts step by step based on the response from the interactive response device 3 and identifies the cause of the abnormal condition.
[0020] [Operation Example 1] Figure 2 is a diagram illustrating the abnormal state information acquired by the cause analysis device 1 in Operation Example 1. In this example, in addition to the alarm name and its identifier, several candidate causes for the alarm are acquired as abnormal state information. Note that the multiple candidate causes may be those listed in the manual, etc.
[0021] Figure 3 illustrates the time-series data acquired by the cause analysis device 1 in operation example 1. In this example, four types of data—position (Pos(X)), position deviation (Err), power supply voltage (Volt), and servo off signal (Signal)—are linked to time (Time) and acquired as time-series data.
[0022] Figure 4 illustrates the prompts generated by the cause analysis device 1 and the responses of the interactive response device 3 in Operation Example 1. First, the cause analysis device 1 presents abnormal state information as input data and generates the first prompt to select time-series data that is related to the abnormal state information within the time-series data and is necessary for investigating the cause of occurrence from multiple candidates.
[0023] In response, the interactive response device 3 indicates that the time-series data required for the investigation are position deviation, servo-off signal, and power supply voltage. The cause analysis device 1 then takes the selected time-series data as input and generates a second prompt to describe the characteristics of the changes in each time-series data.
[0024] In response, the interactive response device 3 indicates that the position deviation has increased since a certain point in time, that there is a change in the servo-off signal, and that the power supply voltage fluctuation is small and within a certain range, as characteristics of the changes in the time-series data. The cause analysis device 1 then uses these characteristics as input to generate a third prompt to present the cause of the abnormal condition along with the reasoning behind it.
[0025] The interactive response device 3 responds to the third prompt by indicating that the cause is most likely to be the input of a servo-off signal. The cause analysis device 1 notifies the user of this response as the investigation result.
[0026] [Example 2] In Example 2, the third prompt has been changed compared to Example 1 in order to make the notification content to the user easier to understand.
[0027] Figure 5 illustrates the prompts generated by the cause analysis device 1 and the responses of the interactive response device 3 in Operation Example 2. In this example, the first prompt and its response, and the second prompt and its response, are the same as in Operation Example 1.
[0028] In the third prompt, the cause analysis device 1 is instructed to display a graph that extracts characteristic parts of the time-series data that served as the basis for presenting the cause of the abnormal condition and its rationale.
[0029] In response, the interactive response device 3 displays a graph including characteristic features of the time-series data of the position deviation and servo-off signal, and then responds with the cause and basis for the abnormal condition.
[0030] [Operation Example 3] In Operation Example 3, the accuracy of identifying the cause of the abnormal condition is improved by increasing the amount of information input to the interactive response device 3. Furthermore, the cause is identified by utilizing feedback from the user.
[0031] Figure 6 illustrates the abnormal state information acquired by the cause analysis device 1 in operation example 3. In this example, similar to operation example 1, the alarm name, its identifier, and several possible causes of the alarm are acquired as abnormal state information. Furthermore, the abnormal state information includes the processing program and the block in which the abnormal state (alarm) occurred.
[0032] Figure 7 illustrates the time-series data acquired by the cause analysis device 1 in operation example 3. In this example, five types of data—position (Pos(X)), position deviation (Err), power supply voltage (Volt), servo off signal (Signal), and program variable (#110)—are linked to time (Time) and acquired as time-series data.
[0033] Figure 8 illustrates the prompts generated by the cause analysis device 1 and the responses of the interactive response device 3 in operation example 3. In the figure, the first to third prompts and responses have some parts omitted that are the same as those in operation example 1.
[0034] First, the cause analysis device 1 presents abnormal state information as input data and then generates the first prompt to select time-series data that is related to the abnormal state information within the time-series data and is necessary for investigating the cause of the occurrence from among multiple candidates.
[0035] In response, the interactive response device 3 indicates that the time-series data required for the investigation are position deviation, servo-off signal, power supply voltage, and variable (#110). The cause analysis device 1 then takes the selected time-series data as input and generates a second prompt to describe the characteristics of the changes in each time-series data.
[0036] In response, the interactive response device 3 indicates that the position deviation has increased since a certain point in time, that there is a change in the servo-off signal, and that the power supply voltage fluctuation is small and within a certain range, as characteristics of the changes in the time-series data. Furthermore, the interactive response device 3 indicates that the value of variable (#110) has changed significantly, and also indicates the meaning of the variable. The cause analysis device 1 then uses these characteristics as input to generate a third prompt to present the cause of the abnormal condition along with the reasoning behind it.
[0037] The interactive response device 3 responds to the third prompt by stating that the cause is most likely the input of a servo-off signal, but here we assume this response was incorrect. Based on the user's feedback that the abnormal condition has not improved, the cause analysis device 1 generates a fourth prompt to prompt the user to suggest other possible causes and their rationale.
[0038] The interactive response device 3 responds to the fourth prompt by indicating, based on the characteristics of the variable change, that an excessive feed rate is likely the cause of the abnormal condition. The cause analysis device 1 notifies the user of this response as the investigation result.
[0039] According to the embodiments illustrated above, the cause analysis device 1 generates a series of stepwise prompts based on abnormal state information and time-series data, thereby obtaining a response from the interactive response device 3 that identifies the cause of the abnormal state. Therefore, the user can reduce the man-hours required for conventional investigations and easily identify the cause of the abnormal state.
[0040] Specifically, the cause analysis device 1 can efficiently and automatically generate the target information by inputting a prompt to the interactive response device 3 at least at any one of the three steps: selecting time-series data related to an abnormal state, extracting the characteristics of the changes in the selected time-series data, and identifying the cause of the abnormal state based on the extracted characteristics.
[0041] In addition, the cause analysis device 1 may present a graph including the characteristic part of the change in the time-series data together with the cause of the abnormal state to be presented to the user and the basis therefor. This makes it easier for the user to specifically understand the basis for the presented cause of the abnormal state.
[0042] The cause analysis device 1 can obtain the data necessary for identifying the cause at the timing when an abnormal state occurs in the control device 2 and proceed with the processing efficiently. At this time, the cause analysis device 1 can improve the identification accuracy by providing the various types of information described above as abnormal state information. Furthermore, the cause analysis device 1 can appropriately extract features by also obtaining the time-series data in the normal state that can be compared with the time-series data in the abnormal state.
[0043] The cause analysis device 1 may generate a prompt so as to present not only the cause of the abnormal state but also a countermeasure method for this abnormal state. Thereby, the user can efficiently obtain useful information when the abnormal state occurs.
[0044] When the cause analysis device 1 presents the cause of the abnormal state, it can accept feedback from the user on the presented content, request the interactive response device 3 to present other causes, and obtain and output a more appropriate response.
[0045] Although embodiments of the present invention have been described in detail above, the present invention is not limited to the individual embodiments described above. These embodiments can be modified in various ways, without departing from the spirit of the invention or without departing from the idea and spirit of the present invention derived from the claims and their equivalents. For example, the order of operations and processes in the embodiments described above are shown as examples only and are not limited thereto. The same applies when numerical values or mathematical formulas are used in the description of the embodiments described above.
[0046] With respect to the embodiments described above, the following additional notes are disclosed. (Addendum 1) A cause analysis device (1) for identifying the cause of an abnormal state in a control device (2), comprising: a data acquisition unit (11) that acquires one or more time-series data linked to time data, including an identifier for representing a unique abnormal state in the control device (2), or abnormal state information including information describing the abnormal state, and information related to control by the control device (2); a prompt generation unit (12) that generates prompts to sequentially instruct an interactive response device (3) to perform each of the stepwise steps for identifying the cause of the abnormal state based on the abnormal state information and the time-series data; a transmission unit (13) that transmits the prompts to the interactive response device (3); a receiving unit (14) that receives a response to the prompts from the interactive response device (3); and a notification unit (15) that notifies the user of the content of the response.
[0047] (Note 2) The cause analysis device (1) described in (Note 1) generates a prompt that commands the interactive response device (3) to perform at least one of the following steps: selecting time-series data related to the abnormal state based on the abnormal state information; presenting the characteristics of the changes in the time-series data; and presenting the cause of the abnormal state based on the abnormal state information and the characteristics of the changes in the time-series data.
[0048] (Supplementary Note 3) The prompt generation unit (12) generates a prompt for instructing the interactive response device (3) to present a graph of the time-series data including the portion where the characteristics of the change have been extracted (Cause analysis device (1) described in Supplementary Note 2).
[0049] (Supplementary Note 4) The data acquisition unit (11) acquires the abnormal state information and the time-series data in response to the occurrence of the abnormal state (Cause analysis device (1) described in any one of Supplementary Notes 1 to 3).
[0050] (Supplementary Note 5) The abnormal state information includes at least any one of the manual information of the control device (2), the inquiry information regarding the control device (2), the source code information of the control device (2), the processing program in which the abnormal state has occurred, the operation history information of the control device (2) performed until the abnormal state has occurred, and the countermeasure methods for each of the conceivable causes of occurrence (Cause analysis device (1) described in any one of Supplementary Notes 1 to 4).
[0051] (Supplementary Note 6) The prompt generation unit (12) generates a prompt for instructing the interactive response device (3) to present a countermeasure method for the cause of occurrence specified for the abnormal state (Cause analysis device (1) described in Supplementary Note 5).
[0052] (Supplementary Note 7) The prompt generation unit (1) receives feedback from the user regarding the correctness of the content of the notified response, and generates a prompt for instructing the interactive response device (3) to present another cause of occurrence different from the response for the abnormal state (Cause analysis device (1) described in any one of Supplementary Notes 1 to 6).
[0053] (Supplementary Note 8) The time-series data includes at least any one of the operation information of the control device (2), the data represented by a physical quantity detectable from the control device (2), the signals input to and output from the control device (2), and the variables used in the processing program of the control device (2) (Cause analysis device (1) described in any one of Supplementary Notes 1 to 7).
[0054] (Note 9) The time-series data includes data of a normal state that can be compared with the abnormal state, as described in any of (Note 1) to (Note 8) of the cause analysis device (1).
[0055] 1. Cause analysis device 2. Control device 3. Interactive response device 10. Control unit 11. Data acquisition unit 12. Prompt generation unit 13. Transmission unit 14. Receiving unit 15. Notification unit 20. Storage unit
Claims
1. A cause analysis device for identifying the cause of an abnormal state in a control device, comprising: a data acquisition unit that acquires one or more time-series data linked to time data, including an identifier for representing a unique abnormal state in the control device, or abnormal state information including information describing the abnormal state, and information related to control by the control device; a prompt generation unit that generates prompts to sequentially instruct an interactive response device to perform each of the stepwise steps for identifying the cause of the abnormal state based on the abnormal state information and the time-series data; a transmission unit that transmits the prompts to the interactive response device; a receiving unit that receives a response to the prompts from the interactive response device; and a notification unit that notifies the user of the content of the response.
2. The cause analysis device according to claim 1, wherein the prompt generation unit generates a prompt that commands the interactive response device to perform at least one of the following steps: selecting the time-series data related to the abnormal state based on the abnormal state information; presenting the characteristics of the changes in the time-series data; and presenting the cause of the abnormal state based on the abnormal state information and the characteristics of the changes in the time-series data.
3. The cause analysis device according to claim 2, wherein the prompt generation unit generates a prompt that instructs the interactive response device to present a graph of the time series data including the portion in which the characteristics of the change have been extracted.
4. The cause analysis device according to any one of claims 1 to 3, wherein the data acquisition unit acquires the abnormal state information and the time-series data in response to the occurrence of the abnormal state.
5. The cause analysis device according to any one of claims 1 to 4, wherein the abnormal state information includes at least one of the following: manual information for the control device, inquiry information regarding the control device, source code information for the control device, the processing program in which the abnormal state occurred, the operation history information of the control device performed up to the occurrence of the abnormal state, and methods for dealing with each of the possible causes.
6. The cause analysis device according to claim 5, wherein the prompt generation unit generates a prompt that instructs the interactive response device to present a method for dealing with the identified cause of the abnormal state.
7. The cause analysis device according to any one of claims 1 to 6, wherein the prompt generation unit receives feedback from the user regarding the correctness of the content of the notified response, and generates a prompt that instructs the interactive response device to present a different cause for the abnormal state than the response.
8. The cause analysis apparatus according to any one of claims 1 to 7, wherein the time-series data includes at least one of the following: operating information of the control device, data represented by a physical quantity detectable from the control device, signals input to and output from the control device, and variables used in the processing program of the control device.
9. The cause analysis apparatus according to any one of claims 1 to 8, wherein the time-series data includes data of a normal state that can be compared with the abnormal state.