State determination apparatus and state determination program
The state determination device and program address the challenge of interpreting two-dimensional graphs by converting equipment status parameters into color-coded time-series images, improving the accuracy of state determination in condition monitoring devices.
Patent Information
- Application Number
- JP2024066783
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-04-17
- Publication Date
- 2025-10-29
AI Technical Summary
Existing condition monitoring devices face challenges in accurately determining the state of equipment due to the difficulty in interpreting two-dimensional graphs, leading to potential inaccuracies in machine learning model labels.
A state determination device and program that convert equipment status parameters into color-coded time-series images, enabling accurate state determination using machine learning models.
Improves the accuracy of state determination by allowing operators and inspectors to easily interpret equipment states through color changes in time-series images, enhancing the reliability of machine learning models.
Smart Images
Figure 2025163484000001_ABST
Abstract
Description
[Technical Field]
[0001] The present disclosure relates to a state determination device and a state determination program that determine the state of a device. [Background technology]
[0002] Patent document 1 describes a condition monitoring device that divides detection information into frames at predetermined time intervals, and generates correct labels by performing machine learning based on two-dimensional image data that represents the distribution of the detection information using a time axis and an axis that indicates the magnitude of the detection information. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] Japanese Patent Application Publication No. 2018-204940 Summary of the Invention [Problem to be solved by the invention]
[0004] In the technology described in Patent Document 1, machine learning is performed using a two-dimensional graph with the magnitude of detection information on the vertical axis and the passage of time on the horizontal axis. Therefore, an inspector inspecting a condition monitoring device must determine whether the labels generated by the machine learning model of the condition monitoring device are appropriate based on the two-dimensional graph used as input. Because a two-dimensional graph represents values corresponding to each time point as points and adjacent points are connected by lines, it can be difficult for an inspector to determine the correct state. For this reason, a worker creating a machine learning model to be installed in a condition monitoring device may be unable to set an appropriate correct label in the machine learning, or an inspector inspecting the machine learning model of a condition monitoring device may be unable to properly determine whether the labels generated by the machine learning model are acceptable or not. This raises concerns about a decline in the accuracy of condition determination by the machine learning model.
[0005] The present disclosure aims to improve the state determination accuracy of a machine learning model that determines a state. [Means for solving the problem]
[0006] One aspect of the present disclosure is a state determination device including a parameter acquisition unit, a color conversion unit, a time-series color image generation unit, and a state determination unit. The parameter acquisition unit is configured to take the device to be judged as the target device, take the state of the target device as the device state, and continuously acquire device state parameters whose values vary depending on the device state.
[0007] The color conversion unit is configured to generate parameter color data indicating a parameter color, which is a color corresponding to a status parameter value, based on a color conversion table in which a correspondence relationship between a status parameter value, which is a value of an equipment status parameter, and a color that is preset according to the status parameter value is set.
[0008] The time series color image generation unit is configured to generate a plurality of pixels to which parameter colors indicated by the parameter color data are set for each of a plurality of parameter color data for the target device, and to generate a time series color image in which the plurality of pixels are arranged in chronological order based on the time at which the device status parameters were acquired.
[0009] The state determination unit is configured to determine, for the target device, the device state corresponding to the time-series color image generated by the time-series color image generation unit, using a machine learning model that has undergone machine learning and inputs a plurality of time-series color images to which labels indicating any of a plurality of device states are attached.
[0010] The state determination device of the present disclosure configured as described above can determine the state of an equipment using a machine learning model that has undergone machine learning and inputs a time-series color image in which multiple pixels, each having a parameter color set corresponding to a state parameter value of the equipment, are arranged in time series. This allows an operator (hereinafter, a model learning operator) who creates the machine learning model to be installed in the state monitoring device and an inspector (hereinafter, a model inspector) who inspects the machine learning model of the condition monitoring device to determine the state of the equipment based on changes in the colors constituting the time-series color image. This prevents the model learning operator from setting an appropriate correct label for the time-series color image or the model inspector from properly determining whether the equipment state determined by the state determination device is good or bad. This allows the state determination device of the present disclosure to improve the state determination accuracy of the machine learning model installed in the state determination device.
[0011] Another aspect of the present disclosure is a state determination program for causing a computer to function as a parameter acquisition unit, a color conversion unit, a time-series color image generation unit, and a state determination unit. A computer controlled by the state determination program of the present disclosure can constitute part of the state determination device of the present disclosure, and can obtain the same effects as the state determination device of the present disclosure. [Brief explanation of the drawings]
[0012] [Figure 1] FIG. 1 is a block diagram showing a configuration of an operating state determination system. [Figure 2] 10 is a flowchart illustrating an operating state determination process. [Figure 3] FIG. 10 is a diagram illustrating the correspondence between power consumption values and colors. [Figure 4] FIG. 2 is a diagram showing the configuration of time-series color image data. [Figure 5] FIG. 10 is a diagram showing the correspondence between operation states, time-series color image data, and operation state labels. [Figure 6] FIG. 10 is a diagram illustrating a correspondence relationship between an operating state and an operating state frame. [Figure 7] FIG. 10 is a diagram illustrating the configuration of an operating status character string. [Figure 8] FIG. 10 is a diagram illustrating a method for determining the operating state of the entire factory. [Figure 9] FIG. 10 shows a compressor power graph and an operating state graph. [Figure 10] FIG. 10 is a diagram showing a time-series operating state image. DETAILED DESCRIPTION OF THE INVENTION
[0013] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. As shown in FIG. 1, the operating state determination system 1 of this embodiment includes an operating state determination device 10, compressors 11, 12, 13, 14, 15, and 16, and current sensors 21, 22, 23, 24, 25, and .
[0014] The operating state determination device 10 includes a display unit 31, an operation input unit 32, a data storage unit 33, a data input / output unit 34, and a control unit 35. The display unit 31 includes a display device (not shown) and displays various images on the display screen of the display device.
[0015] The operation input unit 32 outputs input operation information for identifying an input operation performed by a user via a keyboard and a mouse (not shown). The data storage unit 33 is a storage device for storing various data.
[0016] The data input / output unit 34 inputs and outputs data to and from an external device connected by wire or wirelessly. The control unit 35 is mainly composed of a well-known microcomputer including a CPU 41, a ROM 42, a RAM 43, etc. Various functions of the microcomputer are realized by the CPU 41 executing a program stored in a non-transitory storage medium. In this example, the ROM 42 corresponds to the non-transitory storage medium storing the program. Furthermore, the execution of this program results in the execution of a method corresponding to the program. Note that some or all of the functions executed by the CPU 41 may be configured as hardware using one or more ICs, etc. Furthermore, the number of microcomputers constituting the control unit 35 may be one or more.
[0017] Compressors 11 to 16 are machines that are installed in a factory, for example, and compress air and release the compressed air. The compressors 11, 12, 13, and 14 are constant speed machines in which the rotation speed of the motor that serves as the power source for compressing the air is fixed.
[0018] The compressors 15 and 16 are inverter machines that can change the rotation speed of a motor that serves as a power source for compressing air. The current sensors 21, 22, 23, 24, 25, and 26 detect the values of the currents flowing through the compressors 11, 12, 13, 14, 15, and 16, respectively, and output current data indicating the detection results to the data input / output unit 34 of the operating state determination device 10.
[0019] Next, the procedure of the operation state determination process executed by the control unit 35 will be described. The operation state determination process is executed by starting the operation state determination program 50 stored in the data storage unit 33 through an input operation by the user in order to execute the operation state determination process. The operation state determination program 50 may be pre-installed in the operation state determination device 10, or may be installed from a recording medium or via a network. Examples of recording media include optical disks, magnetic disks, and semiconductor memories.
[0020] The operating state determination process is executed every time a preset data acquisition period (1 second in this embodiment) elapses. Note that the data acquisition period of 1 second is just an example. When the operating state determination process is executed, the CPU 41 of the control unit 35 acquires current data from each of the current sensors 21, 22, 23, 24, 25, and 26 in S10, as shown in FIG.
[0021] In S20, the CPU 41 calculates the power values consumed by the compressors 11, 12, 13, 14, 15, and 16 (hereinafter referred to as power consumption values) based on the current data acquired in S10, and stores power consumption value data indicating the calculated power consumption values in the data storage unit 33. Specifically, the CPU 41, for example, calculates a constant voltage value (e.g., 500 V) to be applied to the motors of the compressors 11 to 16 for each current value indicated by the current data of the current sensors 21 to 26, and stores the power consumption value data indicating the calculated power consumption values in the data storage unit 33. 1 / 2 and the power factor (i.e., 3 1 / 2 × current value × voltage value × power factor) is calculated as the power consumption value of the compressors 11 to 16.
[0022] In S25, the CPU 41 determines whether one minute has passed since the last execution of the process of S30 described below in the operating state determination process. If one minute has not passed since the last execution of the process of S30, the CPU 41 ends the operating state determination process. Note that the one-minute accumulation determination cycle in S25 is an example.
[0023] On the other hand, if one minute has passed since the process of S30 was last executed, the CPU 41 determines in S30 whether or not 15 minutes of power consumption value data has been accumulated for each of the compressors 11 to 16. If 15 minutes of power consumption value data has not been accumulated, the CPU 41 ends the operating state determination process. Note that the accumulated time of 15 minutes is an example.
[0024] On the other hand, if 15 minutes of power consumption value data has been accumulated, the CPU 41 generates power consumption color data in S40, which indicates the power consumption color obtained by converting the power consumption values of the compressors 11 to 16 into colors, for each piece of power consumption value data newly accumulated since the previous processing of S40, based on a color conversion table 60 that presets the correspondence between power consumption values and colors. The color conversion table 60 is stored in the data storage unit 33.
[0025] 3, the color conversion table 60 indicates the correspondence between power consumption values and (R, G, B). (R, G, B) are numerical values in the well-known RGB color system. R represents red, G represents green, and B represents blue, and the R, G, and B values are integers between 0 and 255.
[0026] In this embodiment, the color conversion table 60 sets the correspondence between power consumption values and (R, G, B) so that 0 kW corresponds to dark blue (hereinafter referred to as dark blue) of (R, G, B) = (0, 0, 128), 45 kW corresponds to dark green (hereinafter referred to as dark green) of (R, G, B) = (0, 128, 64), and 90 kW corresponds to light green (hereinafter referred to as light green) of (R, G, B) = (0, 255, 0), and the color gradually changes as the power value increases.
[0027] However, when the power consumption value is equal to or less than a preset lower limit (e.g., 10 kW), the CPU 41 sets the power consumption color to a lower limit color, which is a preset color different from the power consumption color when the power consumption value is greater than the lower limit. In this embodiment, the lower limit color is purple.
[0028] Furthermore, if the CPU 41 is unable to acquire the current data in S10, it generates power consumption color data that indicates a color (hereinafter, "missing data color") that is different from the color set in the color conversion table 60 and the lower limit indicator color and that is preset as a color indicating "missing data." In this embodiment, the missing data indicator color is red.
[0029] As shown in Figure 2, when the processing of S40 is completed, in S50, power consumption color data for the most recent 15 minutes is extracted for each of compressors 11 to 16, and time-series color image data is generated that shows a time-series color image in which the power consumption colors for the most recent 15 minutes are arranged in chronological order.
[0030] 4, specifically, the CPU 41 first generates rectangular pixels for each of the 900 pieces of power consumption color data for 15 minutes, to which the color indicated by the power consumption color data is set, and then arranges the 900 pixels for 15 minutes in time series to form a time-series color image.
[0031] In this embodiment, the CPU 41 arranges 30 pixels horizontally in time series, and when 30 pixels have been arranged horizontally, it turns back to the bottom and arranges the next 30 pixels horizontally, thereby forming a time-series color image in which 30 pixels are arranged horizontally and 30 pixels are arranged vertically. Therefore, the time-series color image is formed so that one second elapses when moving one pixel from left to right, and 30 seconds elapses when moving one pixel from top to bottom.
[0032] As shown in FIG. 2, when the process of S50 is completed, the CPU 41 determines in S60 the operating state of each of the compressors 11 to 16 based on the time-series color image data generated in S50.
[0033] As shown in Figure 5, the operating states are classified into "full load," "wrinkle removal," "hunting," "unload," "stop," and "missing." When there is a lot of light green in the time-series color image data, the operating state is determined to be "full load."
[0034] When light green and dark green are mixed in the time-series color image data and the number of color changes (hereinafter referred to as the number of color changes) is small, the operation state is determined to be "wrinkle removal." When light green and dark green are mixed in the time-series color image data and the number of color changes is large, the operating state is determined to be "hunting."
[0035] In the time-series color image data, if dark blue is mixed with dark green and the color change is gradual, the operating state is determined to be "unloaded." If there is a lot of purple in the time-series color image data, the operating state is determined to be "stopped."
[0036] If there is a lot of red in the time-series color image data, the operating state is determined to be "missing." A machine learning model that determines the operating state by inputting a time-series color image is installed in the operating state determination program 50. In S60, the CPU 41 determines the operating state using this machine learning model.
[0037] Before being installed in the operating status determination program 50, the machine learning model is trained using a large number of time-series color images to which labels indicating correct answers have been added, thereby improving the accuracy of determining the operating status based on the input of time-series color images.
[0038] For example, as shown in Figure 5, time-series color image samples GS1, GS2, GS3, and GS4 are labeled as "full load" and used to train the machine learning model. The time-series color image samples GS11, GS12, GS13, and GS14 are labeled as "wrinkle removal" and used to train the machine learning model.
[0039] The time-series color image samples GS21, GS22, GS23, and GS24 are labeled as "hunting" and used to train the machine learning model. The time-series color image samples GS31, GS32, GS33, and GS34 are labeled as "unloaded" and used to train the machine learning model.
[0040] The time-series color image samples GS41, GS42, GS43, and GS44 are labeled as "stopped" and used to train the machine learning model. The time-series color image samples GS51, GS52, GS53, and GS54 are labeled as "missing" and used to train the machine learning model.
[0041] 2, when the processing of S60 is completed, the CPU 41 adds, in S70, an operation status frame having a preset operation status color representing the determined operation status to each of the time-series color images represented by the determined time-series color image data based on the determination result of S60. The operation status frame is a rectangular frame that surrounds the rectangular time-series color image that was the subject of the determination. Hereinafter, an image in which an operation status frame is added to a time-series color image is referred to as a frame-added color image.
[0042] As shown in FIG. 6, when the operating state is determined to be "full load", an operating state frame FL1 with a light green operating state color is added. If the operation state is determined to be "wrinkle removal," an operation state frame FL2 with a dark green operation state color is added.
[0043] If the operating state is determined to be "hunting", an operating state frame FL3 with an orange operating state color is added. If the operating status is determined to be "unloaded", an operating status frame FL4 with a blue operating status color is added.
[0044] If the operating status is determined to be "stopped", an operating status frame FL5 with a purple operating status color is added. If the operating status is determined to be "missing", an operating status frame FL6 with a red operating status color is added.
[0045] As shown in FIG. 2, when the process of S70 is completed, in S80, the CPU 41 adds an operating state label to each of the time-series color image data that has been judged based on the judgment result in S60.
[0046] As shown in Figure 5, if the operating state is determined to be "full load," an "f" is added to the operating state label. "f" is the initial letter of "FullLoad," which is the English spelling of "full load." However, if "full load" has continued for 30 minutes or more based on past judgment results, an "F" is added to the operating state label instead of "f." "F" is the capital letter of "f."
[0047] If the operation status is determined to be "wrinkle removal," "a" is added to the operation status label. "a" is the initial letter of "Adjust," which is the English spelling of "wrinkle removal." However, if "wrinkle removal" has continued for 30 minutes or more based on past judgment results, "A" is added instead of "a" to the operation status label. "A" is the equivalent of an uppercase "a."
[0048] If the operating state is determined to be "hunting," an "h" is added to the operating state label. "h" is the initial letter of "hunting," which is the English spelling of "hunting." However, if "hunting" has continued for 30 minutes or more based on past judgment results, an "H" is added to the operating state label instead of "h." "H" is the equivalent of an uppercase "h."
[0049] If the operating state is determined to be "unloaded," the letter "u" is added to the operating state label. "u" is the initial letter of "Unload," which is the English spelling of "unload." However, if "unloaded" has continued for 30 minutes or more based on past determination results, the letter "U" is added to the operating state label instead of "u." "U" is the capital letter "u."
[0050] If the operating status is determined to be "stopped," an "s" is added to the operating status label. "s" is the initial letter of "Stop," which is the English spelling of "stopped." However, if, based on past determination results, the "stopped" status has continued for 30 minutes or more, an "S" is added to the operating status label instead of an "s." "S" is the capital letter of "s."
[0051] If the operation status is determined to be "missing," "n" is added as the operation status label. "n" is the initial letter of "NoData," which is the English spelling of "missing." However, if "missing" has continued for 30 minutes or more based on past judgment results, "N" is added as the operation status label instead of "n." "N" is the equivalent of an uppercase "n."
[0052] As shown in FIG. 2, when the process of S80 is completed, the CPU 41 creates an operating state character string in S90 using the operating state label added in S80. The operating status string is created by dividing the operating status labels of compressors 11 to 16 into groups according to compressor type. For example, as shown in Fig. 7, if the operating status labels of compressors 11, 12, 13, 14, 15, and 16 are "S," "H," "U," "F," "S," and "A," respectively, the operating status string will be "SHUF-SA." The "-" in the operating status string is a symbol used to divide the operating status labels into "fixed speed" and "inverter" compressor types and arrange them.
[0053] As shown in FIG. 2, when the process of S90 is completed, the CPU 41 determines the operating status of the entire factory in S100 based on the operating status character string created in S90. As shown in FIG. 8, a regular expression for searching for an operational status character string that satisfies the first condition is, for example, ".*[Nn].*".
[0054] A regular expression for searching for an operational status string that satisfies the second condition is, for example, ".*H.*-.*". A regular expression for searching for an operational status string that satisfies the third condition is, for example, ".*U.*-.*".
[0055] A regular expression for searching for an operational status string that satisfies the fourth condition is, for example, ".*-.*UU.*". A regular expression for searching for an operational status string that satisfies the fifth condition is, for example, ".*A.*-.*A.*".
[0056] A regular expression for searching for an operating status string that satisfies the sixth condition is, for example, ".*h.*-.*". A regular expression for searching for an operating status string that satisfies the seventh condition is, for example, ".*-AF*S*$".
[0057] A regular expression for searching for an operating status string that satisfies the eighth condition is, for example, "^[Ss]*-[Ss]*$". A regular expression for searching for an operating status string that satisfies the ninth condition is, for example, "^[FS]*-.*".
[0058] A regular expression for searching for an operating status string that satisfies the tenth condition is, for example, "*". The priorities of the conditions for searching for the operating status character string are, in descending order, the first condition, the second condition, the third condition, the fourth condition, the fifth condition, the sixth condition, the seventh condition, the eighth condition, the ninth condition, and the tenth condition.
[0059] Each of the first through tenth conditions is classified into one of four operating states: "Good," "Bad," "Unknown," and "Missing." Note that "Missing" refers to a state in which data is missing.
[0060] Therefore, in S100, the CPU 41 determines, in descending order of priority, whether the regular expressions corresponding to the conditions include the operating status strings created in S90. If the regular expressions corresponding to the conditions include the operating status strings, the CPU 41 determines the operating status corresponding to this condition as the operating status of the entire factory. For example, if the regular expression corresponding to the fifth condition includes the operating status string, the operating status of the entire factory is determined to be "poor." If the regular expression corresponding to the seventh condition includes the operating status string, the operating status of the entire factory is determined to be "good."
[0061] As shown in FIG. 2, when the process of S100 ends, the CPU 41 ends the operation state determination. As shown in FIG. 9, the operating state determination device 10 can display a compressor power graph G1 showing the change over time in the power consumption value calculated in S20, and an operating state graph G2 showing the change over time in the operating state determined in S100.
[0062] 10, the operating state determination device 10 can display a time-series operating state image showing the change in the operating state over one day for each of the compressors 11 to 16. The time-series operating state image is generated by arranging time-series color images with operating state frames added (i.e., frame-added color images) in chronological order.
[0063] Specifically, the CPU 41 arranges 60 time-series color images in chronological order at one-minute intervals along the horizontal direction, and when 60 time-series color images have been arranged horizontally, it turns back to the bottom and arranges the next 60 time-series color images along the horizontal direction, thereby forming a time-series operating status image in which 60 time-series color images are arranged horizontally and 24 time-series color images are arranged vertically.
[0064] The operating state determining device 10 configured in this manner is configured to continuously acquire the power consumption values that vary depending on the operating states of the compressors 11-16. The operating status determination device 10 is configured to generate power consumption color data indicating a power consumption color, which is a color corresponding to a power consumption value, based on a color conversion table 60 in which a correspondence relationship between a power consumption value and a color preset according to the power consumption value is set.
[0065] The operating status determination device 10 is configured to generate a plurality of pixels for each of the plurality of power consumption color data in the compressors 11 to 16, with the power consumption color indicated by the power consumption color data set, and to generate a time-series color image in which the plurality of pixels are arranged in chronological order based on the time at which the power consumption was acquired.
[0066] The operating state determination device 10 is configured to use a machine learning model that has undergone machine learning to input multiple time-series color images to which labels indicating one of multiple operating states are attached, and to determine the operating state of compressors 11 to 16 corresponding to the generated time-series color images.
[0067] The operational state determination device 10 can determine the operational state using a machine learning model that has undergone machine learning and inputs a time-series color image in which a plurality of pixels, each having a power consumption color set corresponding to the power consumption value of the compressors 11-16, are arranged in time series. This allows a worker (hereinafter, a model learning worker) who creates the machine learning model to be installed in the operational state determination device 10 and an inspector (hereinafter, a model inspector) who inspects the machine learning model of the operational state determination device 10 to determine the operational state of the compressors 11-16 based on changes in the colors constituting the time-series color image. This prevents the model learning worker from setting an appropriate correct label for the time-series color image, or the model inspector from properly determining whether the operational state determined by the operational state determination device 10 is good or bad. This allows the operational state determination device 10 to improve the state determination accuracy of the machine learning model installed in the operational state determination device 10.
[0068] The power consumption colors are dark blue, dark green, light green, and the like. That is, the power consumption colors are colors other than grayscale. Colored power consumption colors provide more information for expressing differences in power consumption values than grayscale colors, making it easier for model trainers and model inspectors to determine the operating status than when the power consumption colors are grayscale colors. This allows the operating status determination device 10 to improve the accuracy of determining the status of the machine learning model compared to when the power consumption colors are grayscale colors.
[0069] Furthermore, when the power consumption value is equal to or less than a preset lower limit, the operating state determination device 10 is configured to set the power consumption color to a lower limit indicator color, which is a preset color different from the power consumption color when the power consumption value is greater than the lower limit. This makes it easier for the model learning worker and the model checker to recognize that the power consumption value is equal to or less than the lower limit, even when it is not necessary to change the power consumption color when the power consumption value is equal to or less than the lower limit according to the power consumption value.
[0070] The operating state determination device 10 is configured to generate a frame-added color image in which an operating state frame having a preset operating state color indicating the determined operating state is formed to surround a time-series color image corresponding to the determination result, based on the determination result of the operating state. By displaying the generated frame-added color image, the operating state determination device 10 can make it easier for the manager (hereinafter simply referred to as the manager) who manages the compressors 11-16 to recognize the operating states of the compressors 11-16.
[0071] The operating state determination device 10 is configured to determine whether the determined operating state has been maintained for a predetermined continuation determination time (30 minutes in this embodiment) or more. This allows the operating state determination device 10 to notify the administrator whether the determined operating state has been continuing for a long period of time.
[0072] The operational state determination device 10 is configured to determine the operational state of the compressors 11 to 16. The operational state determination device 10 is configured to set a predetermined operational state label using one character to represent the determination result of the operational state for each of the six determined operational states, and to generate an operational state string in which the six operational state labels set for the compressors 11 to 16 are arranged according to a predetermined arrangement condition. The arrangement condition in this embodiment is "divide and arrange the operational state labels of the compressors 11 to 16 by compressor type." In this way, the operational state determination device 10 can determine the pass / fail of the entire facility in which the compressors 11 to 16 are installed using rules such as regular expressions.
[0073] The operational state determination device 10 is configured to use the initial letter of a character string indicating a corresponding operational state as an operational state label. In this way, the operational state determination device 10 allows the administrator to visually recognize the operational state label, thereby enabling the administrator to recognize the content of the operational state corresponding to the operational state label.
[0074] The operating state determination device 10 is configured to change the operating state label according to the duration for which the determined operating state is maintained. In this way, the operating state determination device 10 allows the manager to visually check the operating state label, thereby enabling the manager to recognize whether the operating state corresponding to the operating state label has continued for a long period of time.
[0075] The operational state determination device 10 is configured to use the operational state character string to determine whether the entire factory in which the compressors 11 to 16 are installed is in good condition. In this way, the operational state determination device 10 allows the manager to recognize whether the entire factory in which the compressors 11 to 16 are installed is in good condition.
[0076] In the embodiment described above, the operating state determination device 10 corresponds to a state determination device, the operating state determination program 50 corresponds to a state determination program, the compressors 11 to 16 correspond to target equipment, the operating state corresponds to equipment state, the power consumption corresponds to equipment state parameters, and S10 and S20 correspond to processing as a parameter acquisition unit.
[0077] Furthermore, the power consumption value corresponds to the state parameter value, the power consumption color corresponds to the parameter color, the power consumption color data corresponds to the parameter color data, and S40 corresponds to the processing performed by the color conversion unit.
[0078] Furthermore, S50 corresponds to processing as a time-series color image generation unit, S60 corresponds to processing as a status determination unit, the operating status color corresponds to the equipment status color, S70 corresponds to processing as a frame-added color image generation unit, the operating status label corresponds to the status indication symbol, the operating status character string corresponds to the status symbol array, S100 corresponds to processing as an equipment determination unit, and the equipment corresponds to the equipment.
[0079] Although one embodiment of the present disclosure has been described above, the present disclosure is not limited to the above embodiment and can be implemented in various modifications. [Variation 1] In the above embodiment, the power consumption is used as the device status parameter, but an appropriate device status parameter can be set depending on the device status to be determined. For example, the device status parameter can be current, voltage, temperature, etc.
[0080] [Variation 2] In the above embodiment, the pixel is a color field image, but the pixel may be a grayscale image.
[0081] [Variation 3] In the above embodiment, the operating status label is represented by one character, but the operating status label may be represented by two or more characters, may be represented by symbols only, or may be represented by both characters and symbols. Characters are symbolized words and are a type of symbol. Examples of symbols other than characters include "+" and "x".
[0082] [Variation 4] In the above embodiment, the operating state label is the initial letter of a string representing the corresponding operating state, but the operating state label may be any symbol that can distinguish multiple operating states from one another. For example, the operating state labels for "full load," "wrinkle removal," "hunting," "unload," "stopped," and "missing" may be "1," "2," "3," "4," "5," and "6," or "A," "B," "C," "D," "E," and "F," respectively.
[0083] [Variation 5] In the above embodiment, when the power consumption value is equal to or less than the lower limit, the lower limit indicator color is set as the power consumption color. However, when the power consumption value is equal to or greater than a preset upper limit, the power consumption color may be set to an upper limit indicator color, which is a preset color different from the power consumption color when the power consumption value is less than the upper limit. This makes it easier for the operating state determination device 10 to make the model learning worker and model checker recognize that the power consumption value is equal to or greater than the upper limit, even in cases where it is not necessary to change the power consumption color when the power consumption value is equal to or greater than the upper limit depending on the power consumption value.
[0084] [Variation 6] In the above embodiment, in S30, it is determined whether or not 15 minutes of power consumption color data has been accumulated. However, the process of S30 may be omitted. That is, the processes of S50 and S60 may be executed each time power consumption color data is generated in S40.
[0085] The control unit 35 and the methods described herein may be implemented by a special-purpose computer configured with a processor and memory programmed to perform one or more functions embodied in a computer program. Alternatively, the control unit 35 and the methods described herein may be implemented by a special-purpose computer configured with a processor comprising one or more dedicated hardware logic circuits. Alternatively, the control unit 35 and the methods described herein may be implemented by one or more special-purpose computers configured with a processor and memory programmed to perform one or more functions in combination with a processor configured with one or more hardware logic circuits. Furthermore, the computer program may be stored as instructions executed by a computer on a computer-readable non-transitory tangible storage medium. The methods for implementing the functions of each unit included in the control unit 35 do not necessarily need to include software; all of the functions may be implemented using one or more hardware components.
[0086] In the above embodiments, multiple functions of one component may be realized by multiple components, or one function of one component may be realized by multiple components. Furthermore, multiple functions of multiple components may be realized by one component, or one function realized by multiple components may be realized by one component. Furthermore, part of the configuration of the above embodiments may be omitted. Furthermore, at least part of the configuration of the above embodiments may be added to or substituted for the configuration of another of the above embodiments.
[0087] In addition to the above-described operating state determination device 10, the present disclosure can also be realized in various forms, such as a system including the operating state determination device 10 as a component, a program for causing a computer to function as the operating state determination device 10, a non-transient physical recording medium such as a semiconductor memory on which this program is recorded, and a state determination method. [Technical idea disclosed in this specification] [Item 1] a parameter acquisition unit configured to continuously acquire an apparatus state parameter whose value varies depending on a target apparatus that is a device to be determined and a device state that is a state of the target apparatus; a color conversion unit configured to generate parameter color data indicating a parameter color corresponding to the status parameter value based on a color conversion table in which a correspondence relationship between a status parameter value corresponding to the device status parameter and a color preset according to the status parameter value is set; a time-series color image generating unit configured to generate a plurality of pixels to which the parameter color indicated by the parameter color data is set for each of the plurality of parameter color data for the target device, and to generate a time-series color image in which the plurality of pixels are arranged in chronological order based on the time at which the device status parameters were acquired; a state determination unit configured to determine, for the target device, the device state corresponding to the time-series color image generated by the time-series color image generation unit, using a machine learning model that has undergone machine learning and that receives as input a plurality of the time-series color images to which labels indicating any of the plurality of device states have been added; A state determination device comprising:
[0088] [Item 2] The state determination device according to item 1, A status determination device wherein the parameter color is a color other than grayscale.
[0089] [Item 3] The state determination device according to item 1 or 2, The color conversion unit When the state parameter value is equal to or greater than a predetermined upper limit value, an upper limit indicator color is set as the parameter color, the upper limit indicator color being a color that is set in advance to be different from the parameter color when the state parameter value is less than the upper limit value; and When the state parameter value is equal to or less than a predetermined lower limit value, the state determination device is configured to perform at least one of setting a lower limit indicator color as the parameter color, the lower limit indicator color being a color that is predetermined to be different from the parameter color when the state parameter value is greater than the lower limit value.
[0090] [Item 4] The state determination device according to any one of items 1 to 3, further comprising: a frame-added color image generation unit configured to generate a frame-added color image, based on the determination result by the status determination unit, in which an operating status frame having an equipment status color that is preset as a color indicating the equipment status determined by the status determination unit is formed so as to surround the time-series color image corresponding to the determination result.
[0091] [Item 5] The state determination device according to any one of items 1 to 4, The state determination unit further A state determination device configured to determine whether the determined device state is maintained for a predetermined continuation determination time or longer.
[0092] [Item 6] The state determination device according to any one of items 1 to 4, the state determination unit is configured to determine the device states of the plurality of target devices; a symbol array generation unit configured to set a predetermined state indication symbol using one or more symbols to represent the determination result of the state determination unit for each of the plurality of device states determined by the state determination unit, and to generate a state symbol array in which the plurality of state indication symbols set for each of the plurality of target devices are arranged in accordance with a predetermined arrangement condition.
[0093] [Item 7] Item 5. The state determination device according to item 5, the state determination unit is configured to determine the device states of the plurality of target devices; a symbol array generation unit configured to set a predetermined state indication symbol using one or more symbols to represent the determination result of the state determination unit for each of the plurality of device states determined by the state determination unit, and to generate a state symbol array in which the plurality of state indication symbols set for each of the plurality of target devices are arranged in accordance with a predetermined arrangement condition.
[0094] [Item 8] Item 6. The state determination device according to item 6, The state determination device is configured such that the symbol array generation unit uses the initial letter of a character string indicating the corresponding device state as the state indication symbol.
[0095] [Item 9] The state determination device according to item 6 or 8, The state determination device is configured so that the symbol array generation unit changes the state indication symbol in accordance with the duration for which the device state determined by the state determination unit is maintained.
[0096] [Item 10] Item 7. The state determination device according to item 7, The state determination device is configured so that the symbol array generation unit changes the state indication symbol in accordance with the duration for which the device state determined by the state determination unit is maintained.
[0097] [Item 11] The state determination device according to any one of items 6, 8, and 9, further comprising: A status determination device including an equipment determination unit configured to use the status symbol array to determine whether the entire equipment in which the plurality of target devices are installed is in good condition.
[0098] [Item 12] The state determination device according to item 7 or 10, further comprising: A status determination device including an equipment determination unit configured to use the status symbol array to determine whether the entire equipment in which the plurality of target devices are installed is in good condition.
[0099] [Item 13] Computer, a parameter acquisition unit configured to continuously acquire an equipment status parameter whose value varies depending on a target equipment, the target equipment being a device to be determined, and a device status being a device status; a color conversion unit configured to generate parameter color data indicating a parameter color corresponding to the status parameter value, based on a color conversion table in which a correspondence relationship between a status parameter value corresponding to the device status parameter and a color preset according to the status parameter value is set; a time-series color image generating unit configured to generate a plurality of pixels to which the parameter color indicated by the parameter color data is set for each of the plurality of parameter color data for the target device, and to generate a time-series color image in which the plurality of pixels are arranged in chronological order based on the time at which the device status parameters were acquired; and a state determination unit configured to determine, for the target device, the device state corresponding to the time-series color image generated by the time-series color image generation unit, using a machine learning model that has undergone machine learning and that receives as input a plurality of the time-series color images to which labels indicating any of the plurality of device states have been added. A status determination program to function as a [Explanation of symbols]
[0100] 10...operational state determination device, 11, 12, 13, 14, 15, 16...compressors, 50...operational state determination program, 60...color conversion table
Claims
1. a parameter acquisition unit configured to continuously acquire an apparatus state parameter whose value varies depending on a target apparatus that is a device to be determined and a device state that is a state of the target apparatus; a color conversion unit configured to generate parameter color data indicating a parameter color corresponding to the status parameter value based on a color conversion table in which a correspondence relationship between a status parameter value corresponding to the device status parameter and a color preset according to the status parameter value is set; a time-series color image generating unit configured to generate a plurality of pixels to which the parameter color indicated by the parameter color data is set for each of the plurality of parameter color data for the target device, and to generate a time-series color image in which the plurality of pixels are arranged in chronological order based on the time at which the device status parameters were acquired; a state determination unit configured to determine, for the target device, the device state corresponding to the time-series color image generated by the time-series color image generation unit, using a machine learning model that has undergone machine learning and that receives as input a plurality of the time-series color images to which labels indicating any of the plurality of device states have been added; A state determination device comprising:
2. The state determination device according to claim 1, A status determination device wherein the parameter color is a color other than grayscale.
3. 3. The state determination device according to claim 1 or 2, The color conversion unit When the state parameter value is equal to or greater than a predetermined upper limit value, an upper limit indicator color is set as the parameter color, the upper limit indicator color being a color that is set in advance to be different from the parameter color when the state parameter value is less than the upper limit value; and When the state parameter value is equal to or less than a predetermined lower limit value, the state determination device is configured to perform at least one of setting a lower limit indicator color as the parameter color, the lower limit indicator color being a color that is predetermined to be different from the parameter color when the state parameter value is greater than the lower limit value.
4. 3. The state determination device according to claim 1 or 2, further comprising: a frame-added color image generation unit configured to generate a frame-added color image, based on the determination result by the status determination unit, in which an operating status frame having an equipment status color that is preset as a color indicating the equipment status determined by the status determination unit is formed so as to surround the time-series color image corresponding to the determination result.
5. 3. The state determination device according to claim 1 or 2, The state determination unit further A state determination device configured to determine whether the determined device state is maintained for a predetermined continuation determination time or longer.
6. 3. The state determination device according to claim 1 or 2, the state determination unit is configured to determine the device states of the plurality of target devices; a symbol array generation unit configured to set a predetermined state indication symbol using one or more symbols to represent the determination result of the state determination unit for each of the plurality of device states determined by the state determination unit, and to generate a state symbol array in which the plurality of state indication symbols set for each of the plurality of target devices are arranged in accordance with predetermined arrangement conditions.
7. The state determination device according to claim 5, the state determination unit is configured to determine the device states of the plurality of target devices; a symbol array generation unit configured to set a predetermined state indication symbol using one or more symbols to represent the determination result of the state determination unit for each of the plurality of device states determined by the state determination unit, and to generate a state symbol array in which the plurality of state indication symbols set for each of the plurality of target devices are arranged in accordance with predetermined arrangement conditions.
8. The state determination device according to claim 6, The state determination device is configured such that the symbol array generation unit uses the initial letter of a character string indicating the corresponding device state as the state indication symbol.
9. The state determination device according to claim 6, The state determination device is configured so that the symbol array generation unit changes the state indication symbol in accordance with the duration for which the device state determined by the state determination unit is maintained.
10. The state determination device according to claim 7, The state determination device is configured so that the symbol array generation unit changes the state indication symbol in accordance with the duration for which the device state determined by the state determination unit is maintained.
11. The state determination device according to claim 6, further comprising: A status determination device including an equipment determination unit configured to use the status symbol array to determine whether the entire equipment in which the plurality of target devices are installed is in good condition.
12. The state determination device according to claim 7, further comprising: A status determination device including an equipment determination unit configured to use the status symbol array to determine whether the entire equipment in which the plurality of target devices are installed is in good condition.
13. Computer, a parameter acquisition unit configured to continuously acquire an equipment status parameter whose value varies depending on a target equipment, the target equipment being a device to be determined, and a device status being a device status; a color conversion unit configured to generate parameter color data indicating a parameter color corresponding to the status parameter value, based on a color conversion table in which a correspondence relationship between a status parameter value corresponding to the device status parameter and a color preset according to the status parameter value is set; a time-series color image generating unit configured to generate a plurality of pixels to which the parameter color indicated by the parameter color data is set for each of the plurality of parameter color data for the target device, and to generate a time-series color image in which the plurality of pixels are arranged in chronological order based on the time at which the device status parameters were acquired; and a state determination unit configured to determine, for the target device, the device state corresponding to the time-series color image generated by the time-series color image generation unit, using a machine learning model that has undergone machine learning and that receives as input a plurality of the time-series color images to which labels indicating any of the plurality of device states have been added. A status determination program to function as a
Citation Information
Patent Citations
State monitoring device
JP2018204940A