JP7747251B1 - Factory automation equipment, factory automation system, normal model generation method and program - Google Patents

Abstract

The factory automation equipment (10) according to the present disclosure comprises a housing electrically connected to a frame ground (102), a signal circuit provided inside the housing and having a signal ground (202), a power extraction unit (11) electrically connected to the frame ground (102) and the signal ground (202) to extract power from common mode noise generated between the frame ground (102) and the signal ground (202), and an output unit (12) that outputs information relating to the power extracted from the power extraction unit (11) for display.
(Representative diagram) Figure 1

Description

This disclosure relates to factory automation equipment, etc.

The noise that factory automation equipment is exposed to can be broadly divided into normal mode noise and common mode noise. Common mode noise, in particular, has a greater impact on equipment than normal mode noise, making it important to take measures against it. Patent Document 1 discloses a technology that can detect the power of common mode noise applied to signal lines as a measure against common mode noise.

JP 2009-290668 A

Factory automation equipment is often used alongside many other machines and devices in production sites, and is exposed to a noisy environment caused by the surrounding machines and devices, making it more susceptible to common mode noise. Therefore, it is important to take measures against the common mode noise that factory automation equipment is subjected to. However, because the technology in Patent Document 1 is not applied to factory automation equipment, it is difficult to grasp the common mode noise that factory automation equipment is subjected to.

The purpose of this disclosure is to provide factory automation equipment that can easily grasp the common mode noise to which it is subjected.

The factory automation equipment according to the present disclosure is characterized by comprising a housing electrically connected to a frame ground, a signal circuit provided inside the housing and having a signal ground, a power extraction unit electrically connected to the frame ground and the signal ground to extract power from common mode noise generated between the frame ground and the signal ground, and a model generation unit that generates a normal model for inferring whether or not an abnormality has occurred due to common mode noise from time series data of information relating to the power extracted from the power extraction unit, including information relating to the power when the factory automation equipment is operating stably .

The factory automation equipment disclosed herein can easily grasp the common mode noise to which it is subjected.

FIG. 1 is a diagram showing a functional configuration of factory automation devices in a factory automation system according to a first embodiment of the present disclosure. FIG. 1 is a diagram showing the relationship between a frame ground and a signal ground of a factory automation device according to a first embodiment of the present disclosure. FIG. 1 is a diagram illustrating the relationship between time-series data of information about power during stable operation of a factory automation device according to a first embodiment of the present disclosure and a stable operation lower limit value and a stable operation upper limit value. FIG. 1 is a diagram illustrating the relationship between time-series data of information related to the FA equipment power supply during stable operation of the factory automation equipment according to the first embodiment of the present disclosure and the stable operation lower limit value and the stable operation upper limit value. FIG. 1 is a diagram illustrating an example of a hardware configuration of a factory automation device according to a first embodiment of the present disclosure. FIG. 1 is a diagram illustrating an example of an operation related to a display process of a factory automation device according to the first embodiment of the present disclosure. FIG. 1 is a diagram illustrating an example of an operation related to a learning process of a factory automation device according to a first embodiment of the present disclosure. FIG. 1 is a diagram illustrating an example of an operation related to an inference process for an abnormality occurrence in a factory automation device according to the first embodiment of the present disclosure. FIG. 10 is a diagram illustrating an example of an operation related to a process for comparing occurrence conditions of common mode noise between different installation environments of the FA device 10.

Factory automation equipment (hereinafter also referred to as FA equipment) relating to an embodiment of the present disclosure will be described below with reference to the drawings. Note that identical or corresponding parts in the drawings are designated by the same reference numerals.

Embodiment 1.
The FA device according to the embodiment of the present disclosure is a control device such as a programmable logic controller (PLC) used in the field of factory automation.

As shown in Figures 1 and 2, the factory automation system 1 comprises an FA device 10, a display 20, and an FA device power converter 30. The FA device 10 comprises a housing 101 and a signal circuit 201.

The housing 101 is electrically connected to the frame ground 102. Here, the frame ground is the reference potential of the housing of the device, and the frame ground 102, which is an example of the frame ground, is the reference potential of the housing 101 of the FA device 10. Furthermore, the housing 101 is an example of a housing electrically connected to the frame ground.

Signal circuit 201 is provided inside housing 101 and has signal ground 202. Here, signal ground refers to the reference potential of the signal line, and signal ground 202, which is an example of a signal ground, is the reference potential of signal circuit 201. Furthermore, signal circuit 201 is provided inside the housing and is an example of a signal circuit that has a signal ground.

As shown in Figure 1, in a factory automation system 1 in which a display 20 and an FA equipment power converter 30 are connected to an FA equipment 10, the FA equipment 10 includes a power extraction unit 11, an output unit 12, an FA equipment power information acquisition unit 13, a model generation unit 14, a learning data setting unit 15, an inference unit 16, a memory unit 17, and a communication unit 18.

The power extraction unit 11 is electrically connected to the frame ground 102 and the signal ground 202. The power extraction unit 11 extracts power from common-mode noise generated between the frame ground 102 and the signal ground 202. Common-mode noise refers to noise currents that leak through stray capacitance and return to the power line via the ground. Common-mode noise is characterized by the fact that the direction of the noise currents flowing on the positive and negative sides of the power supply is the same. Because the power extracted by the power extraction unit 11 is extracted from common-mode noise, the greater the extracted power, the greater the common-mode noise at the source. Therefore, the power extracted by the power extraction unit 11 and the magnitude of the common-mode noise at the source are proportional. Therefore, information related to power can also be rephrased as information related to common-mode noise. The power extraction unit 11 is electrically connected to the frame ground and the signal ground and is an example of a power extraction unit that extracts power from common-mode noise generated between the frame ground and the signal ground.

A specific example of the power extraction unit 11 is an energy harvesting module. An energy harvesting module is a module that generates energy by converting electromagnetic wave energy into electrical power. That is, for example, when an energy harvesting module is used as the power extraction unit 11 of the FA device 10 according to an embodiment of the present disclosure, the energy harvesting module is electrically connected to the frame ground 102 and the signal ground 202, and extracts the power by converting common mode noise, which is electromagnetic wave energy generated between the frame ground 102 and the signal ground 202, into electrical power.

The output unit 12 outputs information related to the power extracted from the power extraction unit 11 for display. More specifically, the output unit 12 extracts the information related to the power from the power extraction unit 11 and outputs it to the display 20 connected to the FA equipment 10 via the communication unit 18. The information related to the power output from the output unit 12 is displayed on the display 20. Here, the information related to the power includes at least the value of the power extracted from the power extraction unit 11 (power value). The information related to the power displayed on the display 20 may be time-series data. For example, the display 20 may display time-series data of the power value, as shown in FIG. 3, and may further display the lower and upper stable operation limits. For example, by checking the range from the minimum to maximum power values at each time displayed on the display 20, the user can visually grasp the value of the power extracted from the power extraction unit 11, i.e., the magnitude of the common-mode noise received by the FA equipment 10. Furthermore, by checking whether the value of the power extracted from the power extraction unit 11 falls below the stable operation lower limit or exceeds the stable operation upper limit, it becomes possible to visually grasp the timing at which common mode noise received by the FA equipment 10 occurs. The stable operation lower limit and stable operation upper limit will be described later in relation to the model generation unit 14. Here, the display unit 20 is an external device, such as a computer such as a personal computer or a programmable display. Note that the FA equipment 10 may have a display unit (not shown) that displays the output from the output unit 12. The output unit 12 is an example of an output unit that outputs information related to the power extracted from the power extraction unit for display.

The FA equipment power supply information acquisition unit 13 acquires time-series data of information related to the FA equipment power supply. Here, the FA equipment power supply refers to a DC power supply that supplies the DC voltage used by the FA equipment 10. In the FA equipment 10, AC voltage supplied from an external power supply (not shown) is converted to DC voltage by an FA equipment power converter 30 externally attached to the FA equipment 10, and the converted DC voltage is supplied from the FA equipment power converter 30 to the FA equipment 10. Therefore, the information related to the FA equipment power supply is information related to the DC power supplied from the FA equipment power converter 30 connected to the FA equipment 10, and includes, for example, the voltage value of the DC power supplied from the FA equipment power converter 30 connected to the FA equipment 10. Note that although the FA equipment power converter 30 has been described as a separate device external to the FA equipment 10, it may also be built into the FA equipment 10. Alternatively, the FA equipment 10 may be one of the constituent units, such as a power supply unit that is one of the units of a PLC that supplies power when the FA equipment 10 is a building block type PLC. While the FA equipment power supply converter 30 has been described as converting AC voltage to DC voltage, the present invention is not limited to this. For example, the FA equipment power supply converter 30 may convert a DC voltage to a smaller voltage without changing the DC voltage, such as converting a 24 V DC voltage to a 5 V DC voltage. The FA equipment power supply information acquiring unit 13 is an example of an FA equipment power supply information acquiring unit that acquires time-series data of information related to the FA equipment power supply.

The model generation unit 14 generates a trained model for inferring the presence or absence of an abnormality due to common mode noise from time-series data on power information. In particular, the model generation unit 14 generates a normal model as a trained model for inferring the presence or absence of an abnormality due to common mode noise using training data including time-series data on power information during stable operation of the FA equipment 10. Hereinafter, the model generation unit 14 will be described as generating a normal model as a trained model. Note that, to generate the trained model, data from the time-series data on power information during stable operation of the FA equipment 10, acquired over a fixed acquisition period set by the user, is used as training data. The acquisition period for the training data will be described later in the training data setting unit 15. The model generation unit 14 is an example of a model generation unit that generates a trained model for inferring the presence or absence of an abnormality due to common mode noise from time-series data on power information.

The learning algorithm used by the model generation unit 14 can be any known algorithm, such as supervised learning or unsupervised learning. As an example, we will explain the application of k-means (clustering), an unsupervised learning method.

The model generation unit 14 learns time series data regarding power information when there are no abnormalities due to common mode noise, for example, using a grouping method based on the k-means method, through so-called unsupervised learning.

K-means is a non-hierarchical clustering algorithm that uses the cluster mean to classify a given number of clusters into k.

Specifically, the k-means algorithm is processed as follows: First, a cluster is randomly assigned to each data x i . Next, the center V j of each cluster is calculated based on the assigned data. Next, the distance between each x i and each V j is calculated, and x i is reassigned to the cluster with the closest center. If the cluster assignment for all x i has not changed in the above process, or if the amount of change is below a certain threshold set in advance, it is determined that convergence has occurred and the process ends.

In embodiment 1 of the present disclosure, time series data regarding power when no abnormalities due to common mode noise occur is learned by so-called unsupervised learning, in accordance with learning data created based on time series data regarding power.

In embodiment 1 of the present disclosure, a specific method in which the model generation unit 14 generates a normal model as a trained model will be described with reference to Figure 3. As shown in Figure 3, time-series data regarding power is shown by a solid line, with power on the vertical axis and time on the horizontal axis. This time-series data regarding power is learning data during stable operation of the FA equipment 1, and Figure 3 illustrates data for a portion (a single cycle) of the acquisition period of the periodic processing.

In embodiment 1 of the present disclosure, the trained model is a model for classifying, for example, data into clusters indicating stable operation. As shown in Figure 3, the stable operation lower limit and stable operation upper limit are indicated by dotted lines, and the area surrounded by the stable operation lower limit and stable operation upper limit corresponds to the cluster indicating stable operation. In other words, if the time series data regarding power does not belong to a cluster indicating stable operation, it is inferred that an abnormality has occurred. In this way, in embodiment 1 of the present disclosure, time series data regarding power is input into the trained model, and the obtained time series data regarding power is classified into a cluster indicating stable operation, thereby inferring whether an abnormality due to common mode noise has occurred.

The model generation unit 14, for example, repeatedly executes periodic processing during stable operation and aligns the start and end points of each cycle to generate clusters representing stable operation in which power-related information (power values) is distributed. The model generation unit 14 then calculates the smallest value among the values belonging to the cluster representing stable operation at each time as the stable operation lower limit, and calculates the largest value among the values belonging to the cluster representing stable operation at each time as the stable operation upper limit. Furthermore, instead of inferring the occurrence of an abnormality based on which cluster the time-series data of power-related information belongs to, as described above, the occurrence of an abnormality may be inferred based on whether the time-series data of power-related information falls outside a threshold value, such as the stable operation lower limit or the stable operation upper limit. That is, an abnormality may be inferred when the time-series data of power-related information falls below the stable operation lower limit or when the time-series data of power-related information exceeds the stable operation upper limit.

As shown in FIG. 4, the time-series data on the FA equipment power supply, with the vertical axis representing voltage and the horizontal axis representing time, is shown by a solid line. Similarly to FIG. 3, the stable operation lower limit and the stable operation upper limit are also shown by dotted lines. The model generation unit 14 may acquire time-series data on the FA equipment power supply during stable operation of the FA equipment 10, as shown in FIG. 4, multiple times during repeated execution of periodic processing during stable operation, in addition to time-series data on the power supply during stable operation of the FA equipment 10, as shown in FIG. 3. By aligning the start and end points of the time width of each period, the model generation unit 14 may generate a normal model by using a cluster representing stable operation in which the information (voltage values) on the FA equipment power supply is distributed as a trained model. That is, the model generation unit 14 may generate the normal model using the training data used to generate the normal model as time-series data on the FA equipment power supply during stable operation of the FA equipment 10, as shown in FIG. 4, in addition to time-series data on the power supply during stable operation of the FA equipment 10, as shown in FIG. 3. Note that a normal model using both the time-series data of information related to power during stable operation of the FA device 10 and the time-series data of information related to the FA device power supply during stable operation of the FA device 10 becomes a three-dimensional cluster (for example, a cluster with time on the X-axis, voltage on the Y-axis, and power on the Z-axis) that combines FIGS. 3 and 4 .

The learning data setting unit 15 sets the learning data acquisition period. The learning data setting unit 15 may set the learning data acquisition period to, for example, a period determined by the user as being in a stable operating state for the FA device 10, or may set the learning data acquisition period to a predetermined period starting from a certain time after the FA device 10 has been started up until stable operation has been achieved if it is clear that the FA device 10 is in a stable operating state, such as when it is clear that the installation environment of the FA device 10 is not a noisy environment.

The inference unit 16 infers whether an abnormality has occurred using time-series data on power information and the trained model. More specifically, the inference unit 16 inputs the time-series data on power information acquired by the power extraction unit 11 into the trained model, and outputs whether an abnormality has occurred as inferred from the time-series data on power information. Here, a normal model is used as the trained model, and the inference unit 16 inputs the time-series data on power information into the trained model to infer whether the FA equipment 10 is in a stable operating state from the time-series data on power information. Here, if the inference unit 16 infers that the FA equipment 10 is in a stable operating state, this means that no abnormality has occurred. If the inference unit 16 infers that the FA equipment 10 is not in a stable operating state, this means that an abnormality has occurred. In other words, the inference unit 16 inputs the time-series data on power information into the trained model, and infers whether the FA equipment 10 is in a stable operating state from the time-series data on power information, thereby inferring whether an abnormality has occurred. More specifically, as described above, the inference unit 16 inputs time-series data regarding power into the trained model, and infers whether or not an abnormality has occurred due to common-mode noise by classifying the obtained time-series data regarding power into a cluster indicating stable operation. The inference unit 16 is an example of an inference unit that infers whether or not an abnormality has occurred using the time-series data regarding power and the trained model.

Here, the inference unit 16 may infer whether or not an abnormality has occurred using time series data on information related to power, time series data on information related to the FA equipment power supply, and the trained model. That is, the data used by the inference unit 16 may include time series data on information related to power as well as time series data on information related to the FA equipment power supply. Furthermore, hereinafter, the time series data on information related to power and the time series data on information related to the FA equipment power supply may collectively be referred to simply as time series data.

When the inference unit 16 infers that an abnormality has occurred, the memory unit 17 stores the time the abnormality occurred, information about the power at the time the abnormality occurred, and information about the FA equipment power supply at the time the abnormality occurred. That is, when the inference unit 16 infers that an abnormality has occurred, the memory unit 17 stores the power value and voltage value at the time the abnormality occurred. Note that the time the abnormality occurred may be the time when the inference unit 16 infers that an abnormality has occurred, or the time when the time-series data that the inference unit 16 infers that an abnormality has occurred is acquired. The memory unit 17 is an example of a memory unit that, when the inference unit infers that an abnormality has occurred, stores the time the abnormality occurred and at least one of information about the power at the time the abnormality occurred and information about the FA equipment power supply at the time the abnormality occurred.

The memory unit 17 also stores the trained model generated by the model generation unit 14. Note that the trained model stored in the memory unit 17 does not have to be one generated by the model generation unit 14; for example, a trained model trained by a learning device other than the FA equipment 10 may be stored. Furthermore, the inference unit 16 uses the trained model by referring to the memory unit 17.

The communication unit 18 communicates with the display device 20 connected to the FA device 10. For example, the FA device 10 can connect to the display device 20 via the communication unit 18 and exchange data such as time series data with the display device 20, thereby displaying the time series data on the display device 20.

An example of the hardware configuration of the FA device 10 will be described with reference to Figure 5. The FA device 10 includes a non-volatile memory 1001, a volatile memory 1002, a communication interface 1003, and a processor 1004, all of which are connected via a bus 1000.

The non-volatile memory 1001 includes, for example, a ROM (Read Only Memory). The non-volatile memory 1001 stores programs executed by the processor 1004.

Volatile memory 1002 includes, for example, RAM (Random Access Memory). Volatile memory 1002 stores programs that processor 1004 has stored from non-volatile memory 1001. Volatile memory 1002 also functions as working memory when processor 1004 executes programs.

In addition, non-volatile memory 1001 and volatile memory 1002 function as memory unit 17.

The communication interface 1003 includes, for example, an I/O (Input/Output) interface, a USB (Universal Serial Bus) port, and a serial port. The FA device 10 communicates with the display device 20 via the communication interface 1003, and exchanges data such as time-series data with the display device 20. The communication interface 1003 functions as the communication unit 18.

The processor 1004 includes, for example, a CPU and an MPU (Micro Processing Unit). The processor 1004 stores a program from the non-volatile memory 1001 to the volatile memory 1002 and executes the program stored in the volatile memory 1002, thereby realizing the output unit 12, the FA equipment power supply information acquisition unit 13, the model generation unit 14, the learning data setting unit 15, and the inference unit 16 and performing each operation. Each operation will be described later in the explanation of the operation of the FA equipment 10. The program executed by the processor 1004 may also be realized by a computer program product, which is a recording medium on which the program is stored. An example of the recording medium in this case is a non-transitory computer-readable medium on which the program is stored.

Note that non-volatile memory 1001, volatile memory 1002, and processor 1004 are examples of signal circuits having a signal ground.

Next, the operation of the FA device 10 configured in this manner will be described. The FA device 10 performs the following operations, which will be described later: operation related to the display processing of the FA device 10; operation related to the learning processing of the FA device 10; operation related to the inference processing of the occurrence of an abnormality in the FA device 10; and operation related to the comparison processing of the occurrence status of common mode noise between different installation environments of the FA device 10. An example of the operation related to the display processing of the FA device 10 will be described with reference to FIG. 6. Here, FIG. 6 is a flowchart showing an example of the operation related to the display processing of the FA device 10, and describes the display processing of the FA device 10 when displaying on the display device 20 connected to the FA device 10.

As shown in Figure 6, the power extraction unit 11 extracts power from the common mode noise generated between the frame ground 102 and the signal ground 202 (step S11).

After the power extraction unit 11 extracts power from the common mode noise generated between the frame ground 102 and the signal ground 202 in step S11, the output unit 12 outputs information about the power extracted from the power extraction unit 11 to the display 20 connected to the FA device 10 (step S12). After step S12, the display 20 connected to the FA device 10 displays information about the power extracted from the power extraction unit 11 as shown in Figure 3, and the FA device 10 ends its operation related to the display processing of the FA device 10. Note that information about the FA device power supply when the FA device 10 is operating stably may also be displayed as shown in Figure 4.

Next, an example of the operation related to the learning process of the FA device 10 configured as described above will be described with reference to Figure 7. Here, Figure 7 is a flowchart showing an example of the operation related to the learning process of the FA device 10.

As shown in Figure 7, the learning data setting unit 15 sets the period for acquiring learning data (step S21).

After the learning data setting unit 15 sets the learning data acquisition period in step S21, the learning data setting unit 15 acquires learning data during the set acquisition period (step S22).

After acquiring the learning data for the acquisition period set by the learning data setting unit 15 in step S22, the model generation unit 14 learns the time series data when no abnormalities due to common mode noise occur, using so-called unsupervised learning, according to the learning data acquired from the learning data setting unit 15, and generates a learned model (step S23).

In step S23, the model generation unit 14 uses the learning data acquired from the learning data setting unit 15 to generate a trained model for inferring whether or not an abnormality due to common mode noise has occurred. After that, the model generation unit 14 stores the generated trained model in the storage unit 17 (step S24). After step S24, the FA device 10 terminates operations related to the learning process of the FA device 10.

Next, an example of the operation related to the inference processing of the FA device 10 configured as described above will be described with reference to Figure 8. Here, Figure 8 is a flowchart showing an example of the operation related to the inference processing of the FA device 10.

As shown in Figure 8, the power extraction unit 11 and the FA equipment power supply information acquisition unit 13 acquire time series data (step S31).

After the power extraction unit 11 and the FA equipment power supply information acquisition unit 13 acquire time series data in step S31, the inference unit 16 inputs the time series data acquired by the power extraction unit 11 and the FA equipment power supply information acquisition unit 13 into the learned model generated by the model generation unit 14 (step S32).

In step S32, the inference unit 16 inputs the time series data acquired by the power extraction unit 11 and the FA equipment power supply information acquisition unit 13 into the learned model generated by the model generation unit 14. Then, the inference unit 16 acquires the learned model generated by the model generation unit 14 from the memory unit 17, inputs the time series data acquired by the power extraction unit 11 and the FA equipment power supply information acquisition unit 13 into the acquired learned model, and infers whether or not an abnormality has occurred (step S33).

If the inference unit 16 infers that no abnormality has occurred (step S33: No), the process returns to step S31, where the power extraction unit 11 and the FA equipment power supply information acquisition unit 13 acquire new time series data again in step S31, and the FA equipment 10 then repeats the operations from step S32.

If the inference unit 16 infers that an abnormality has occurred (step S33: Yes), the memory unit 17 stores the power value and voltage value at the time the abnormality occurred (step S34). After step S34, the FA device 10 terminates operations related to the inference process of the FA device 10.

Next, an example of the operation relating to the process of comparing the occurrence status of common mode noise between different installation environments of the FA device 10 configured as described above will be described with reference to Figure 9. Here, Figure 9 is a flowchart showing an example of the operation relating to the process of comparing the occurrence status of common mode noise between different installation environments of the FA device 10 (installation environment A, installation environment B). Specifically, normal models are created between the different installation environments of the FA device 10 (installation environment A, installation environment B), and feature quantities calculated from each normal model are compared.

As shown in Figure 9, the model generation unit 14 and learning data setting unit 15 of the FA device 10 perform learning processing in installation environment A (step S41). That is, the FA device 10 performs learning processing from step S21 to step S24 shown in Figure 7 in installation environment A, and generates a normal model in installation environment A as a trained model in installation environment A.

After the model generation unit 14 and learning data setting unit 15 of the FA device 10 perform learning processing for installation environment A in step S41, the inference unit 16 of the FA device 10 calculates the amount of power in installation environment A as a feature of the normal model for installation environment A (step S42). In order to understand the occurrence status of common mode noise in installation environment A, it is necessary to calculate the total amount of power generated in installation environment A, i.e., the amount of power in installation environment A. Therefore, using the generated normal model for installation environment A, the FA device 10 calculates the "area surrounded by the stable operation lower limit and stable operation upper limit (i.e., the area surrounded by the stable operation lower limit and stable operation upper limit and the start and end points of the time span in installation environment A in Figure 3)" in the normal model for installation environment A as a feature of the normal model for installation environment A.

After the inference unit 16 of the FA equipment 1 calculates the features of the normal model in installation environment A in step S42, the model generation unit 14 and learning data setting unit 15 of the FA equipment 10 perform learning processing in installation environment B (step S43).

After the model generation unit 14 and learning data setting unit 15 of the FA equipment 10 perform the learning process for installation environment B in step S43, the inference unit 16 of the FA equipment 10 calculates the amount of power in installation environment B as a feature of the normal model for installation environment B (step S44). That is, using the generated normal model for installation environment B, the inference unit 16 of the FA equipment 10 calculates the "area surrounded by the stable operation lower limit and the stable operation upper limit (i.e., the area surrounded by the stable operation lower limit and the stable operation upper limit and the start and end points of the time range in installation environment B in Figure 3)" in the normal model for installation environment B as a feature of the normal model for installation environment B.

After the inference unit 16 of the FA device 10 calculates the feature quantities of the normal model in installation environment B in step S44, the inference unit 16 of the FA device 10 compares the feature quantities of the normal models for installation environments A and B by calculating the change in the feature quantities of the normal model due to the change in the installation environment (step S45). More specifically, the inference unit 16 of the FA device 10 calculates the difference between the amount of power consumed in installation environment B and the amount of power consumed in installation environment A as the change in the amount of power consumed due to the change in the installation environment. Therefore, if the change in the amount of power consumed due to the change in the installation environment is a positive value, it indicates that the amount of power consumed, i.e., the common-mode noise, increased due to the change in the installation environment (the change in the installation environment was ineffective as a noise countermeasure, and in fact worsened). If it is 0, it indicates that the common-mode noise did not change even after the change in the installation environment (the change in the installation environment was ineffective as a noise countermeasure). If it is a negative value, it indicates that the common-mode noise decreased due to the change in the installation environment (the change in the installation environment was effective as a noise countermeasure).

After the inference unit 16 of the FA device 10 calculates the amount of change in the feature quantities of the normal model due to the change in the installation environment in step S45, the output unit 12 outputs the amount of change for display as a comparison result of the feature quantities of the normal model before and after the change in the installation environment in step S46. The inference unit 16 of the FA device 10 then terminates the operation related to the process of comparing the occurrence status of common mode noise between different installation environments of the FA device 10.

In the above example, the change in power consumption before and after the change in the installation environment was calculated as a feature of the normal model before and after the change in the installation environment. However, as a variant, the rate of change in power consumption before and after the change in the installation environment may be calculated and compared. Specifically, the rate of change in power consumption may be calculated by subtracting the power consumption in installation environment A from the power consumption in installation environment B and dividing the difference by the power consumption in installation environment A, and the rate of change may be output for display.

As a variation of the above example, the upper and lower ranges of the stable operation lower and upper limits of power before and after the change in the installation environment may be calculated as features of the normal model and compared before and after the change in the installation environment. Specifically, the difference between the upper and lower ranges of the stable operation lower and upper limits at a specific time in installation environment B and the upper and lower ranges of the stable operation lower and upper limits at a specific time in installation environment A is calculated as the change in the range. A positive change in the range indicates an increase in common-mode noise due to the change in the installation environment. A change of zero indicates no change in common-mode noise due to the change in the installation environment. A negative change in the range indicates a decrease in common-mode noise due to the change in the installation environment. Note that a specific time refers to the same time within one cycle, and the upper and lower ranges of the stable operation lower and upper limits at the same time within one cycle in installation environments A and B are compared.

Furthermore, as a variation of the above example, when generating a three-dimensional normal cluster that takes into account information about the FA equipment power supply before and after the change in the installation environment in addition to information about power before and after the change in the installation environment, the volume of the part surrounded by the stable operation lower limit and stable operation upper limit of the information about power (power) and the information about the FA equipment power supply (voltage) in installation environments A and B and the start and end points of the time span may be calculated and compared as feature quantities of the normal model before and after the change in the installation environment. Specifically, the inference unit 16 of the FA equipment 10 calculates the difference between the volume in installation environment B minus the volume in installation environment A as the amount of change in common-mode noise due to the change in the installation environment. Therefore, if the amount of change in volume due to a change in the installation environment is a positive value, it indicates that the common mode noise increased due to the change in the installation environment (the change in the installation environment had no effect on noise countermeasures, and in fact worsened), if it is 0, it indicates that the common mode noise did not change even after the change in the installation environment (the change in the installation environment had no effect on noise countermeasures), and if it is a negative value, it indicates that the common mode noise decreased due to the change in the installation environment (the change in the installation environment had an effect on noise countermeasures).In addition, instead of the amount of change in volume before and after a change in the installation environment, the rate of change in volume before and after a change in the installation environment may be calculated.

In this way, the FA device 10 according to this embodiment extracts power from the common mode noise generated between the frame ground 102 and the signal ground 202, and outputs information about the extracted power from the power extraction unit 11 for display. If it is possible to output information about power, it is possible to display information about power, for example, as shown in FIG. 3, at the output destination to which the information about power is output. This makes it possible to display information about power, i.e., information about the common mode noise received by the FA device 10, which has the effect of making it easy to understand the common mode noise received by the FA device 10 itself.

Furthermore, by being able to display information about the common mode noise received by the FA equipment 10, for example, if a noise countermeasure is implemented by changing the installation environment of the FA equipment 10, simply by displaying information about the common mode noise received by the FA equipment 10 in each installation environment and comparing the range of minimum to maximum power values at each time in each of the displayed installation environments, it becomes possible to visually grasp the magnitude of the common mode noise received by the FA equipment 10 in each installation environment and verify the noise countermeasure. This eliminates the need for noise countermeasure verification work and the preparations required for the verification work, thereby achieving the effect of reducing the effort and time required for the noise countermeasure verification work.

Furthermore, the FA device 10 according to this embodiment generates a trained model for inferring the presence or absence of an abnormality due to common mode noise from time-series data on information related to power. It is possible to generate a trained model that can infer the presence or absence of an abnormality due to common mode noise by learning common mode noise patterns based on time-series data on information related to common mode noise. This provides the advantage of being able to obtain an index for detecting an abnormal state due to common mode noise in the installation environment of the FA device. In particular, the FA device 10 according to the present disclosure generates a normal model as a trained model from time-series data on information related to power during stable operation of the FA device 10. This provides the advantage of being able to obtain an index for detecting an abnormal state due to common mode noise based on time-series data on information related to common mode noise during stable operation of the FA device 10.

Furthermore, when the installation environment is changed, a normal model is generated for each installation environment, and the generated normal models are used to calculate and compare the features of the normal models before and after the installation environment change, making it possible to compare the common mode noise occurrence conditions between each installation environment. Therefore, the effectiveness of noise countermeasures resulting from the installation environment change can be understood from the change in the common mode noise occurrence conditions due to the installation environment change, and an index can be obtained to evaluate the effectiveness of noise countermeasure improvements.

Furthermore, in the FA equipment 10 according to this embodiment, a normal model is generated as a trained model from time series data on information related to the FA equipment power supply during stable operation of the FA equipment 10, in addition to time series data on information related to the power supply during stable operation of the FA equipment 10. In this way, by acquiring time series data on information related to the FA equipment power supply as training data in addition to time series data on information related to power, the number of parameters used to generate the normal model increases, which has the effect of enabling the accuracy of the normal model to be improved.

Furthermore, the FA device 10 according to this embodiment uses time-series data on information related to power and a trained model to infer whether an abnormality has occurred. Therefore, when acquiring time-series data on information related to power, i.e., time-series data on information related to common-mode noise received by the FA device 10, it is possible to automatically detect the occurrence of an abnormality in common-mode noise.

Furthermore, the FA equipment 10 according to this embodiment infers whether an abnormality has occurred using time-series data on information related to power, time-series data on information related to the FA equipment power supply, and the trained model. In this way, by inputting time-series data on information related to the FA equipment power supply, in addition to time-series data on information related to power, into the trained model, the number of parameters used by the inference unit 16 increases, improving the accuracy of inference and enabling improved accuracy in detecting abnormalities.

In addition, in the FA device 10 according to this embodiment, when the inference unit 16 infers that an abnormality has occurred, the memory unit 17 stores the time the abnormality occurred and at least one of information about the power at the time the abnormality occurred and information about the FA device power supply at the time the abnormality occurred. By recording detailed information about the occurrence of an abnormality in this way, it is possible to quickly identify the cause and take countermeasures even when an abnormality occurs.

Here, the output unit 12 has been described as outputting information related to the power extracted from the power extraction unit 11 for display. However, it may also output information related to the FA equipment power source acquired by the FA equipment power source information acquisition unit 13 for display. More specifically, the output unit 12 may acquire information related to the FA equipment power source from the FA equipment power source information acquisition unit 13 and output the information to a display 20 connected to the FA equipment 10 via the communication unit 18, thereby displaying the information related to the FA equipment power source output from the output unit 12 on the display 20. The information related to the FA equipment power source displayed on the display 20 may also be time-series data. For example, in addition to the time-series data related to power as shown in FIG. 3, the display 20 may also display time-series data of the FA equipment power source value as shown in FIG. 4. Furthermore, the stable operation lower limit and the stable operation upper limit may also be displayed. Even in such a case, the user can visually grasp the voltage value extracted from the FA equipment power supply converter 30 by checking, for example, the range from the minimum value to the maximum value of the voltage value at each time or the range from the lower stable operation limit value to the upper stable operation limit value, which are displayed on the display device 20. Furthermore, since it is now possible to display time series data of information about the FA equipment power supply in addition to time series data of information about power, it becomes easier to grasp the relationship between the time series data of information about power and the time series data of information about the FA equipment power supply, which also has the effect of facilitating troubleshooting when an abnormality occurs.

The present disclosure allows various embodiments and modifications without departing from the broad spirit and scope of the present disclosure. Furthermore, the above-described embodiments are intended to explain the present disclosure and do not limit the scope of the present disclosure. In other words, the scope of the present disclosure is defined by the claims, not the embodiments. Various modifications made within the scope of the claims and the meaning of equivalent disclosures are considered to be within the scope of the present disclosure.

1. Factory automation system,
10 Factory automation equipment, 11 Power extraction section, 12 Output section,
13 FA equipment power supply information acquisition unit, 14 Model generation unit, 15 Learning data setting unit,
16 inference unit, 17 storage unit, 18 communication unit, 20 display unit, 30 FA equipment power converter,
101 housing, 102 frame ground, 201 signal circuit,
202 signal ground, 1000 bus, 1001 non-volatile memory,
1002 Volatile memory, 1003 Communication interface, 1004 Processor

Claims (11)

a housing electrically connected to a frame ground;
a signal circuit provided inside the housing and having a signal ground;
a power extraction unit electrically connected to the frame ground and the signal ground, extracting power from common mode noise generated between the frame ground and the signal ground;
A factory automation device comprising:
a model generation unit that generates a normal model for inferring whether or not an abnormality due to the common mode noise has occurred from time-series data of information about the power extracted from the power extraction unit, the information being about the power when the factory automation equipment is operating stably; and
1. A factory automation device comprising: an inference unit that infers whether or not the abnormality has occurred using time-series data of information related to the power and the normal model;
10. The factory automation equipment of claim 1, further comprising: an FA equipment power supply information acquisition unit that acquires time-series data of information related to the FA equipment power supply;
2. The factory automation device according to claim 1, wherein the model generation unit generates the normal model from time series data of information related to the FA device power supply during the stable operation in addition to time series data of information related to power during the stable operation. an FA equipment power supply information acquisition unit that acquires time-series data of information related to the FA equipment power supply;
the model generation unit generates the normal model from time series data of information about the FA equipment power supply during the stable operation in addition to time series data of information about the power during the stable operation,
3. The factory automation device according to claim 2, wherein the inference unit infers whether or not the abnormality has occurred by using time-series data of the information related to the power, time-series data of the information related to the FA device power supply, and the normal model. a storage unit configured to store, when the inference unit infers that an abnormality has occurred, a time when the abnormality occurred and at least one of information about the power at the time when the abnormality occurred and information about the FA equipment power supply at the time when the abnormality occurred;
5. The factory automation equipment of claim 4, further comprising: an output unit that outputs information about the power extracted from the power extraction unit for display;
6. The factory automation device according to claim 1, further comprising: an output unit that outputs , for displaying, information about the power extracted from the power extraction unit and information about the FA equipment power supply acquired from the FA equipment power supply information acquisition unit;
6. The factory automation device according to claim 3, further comprising : an output unit that outputs information about the power extracted from the power extraction unit for display;
Furthermore,
the model generation unit generates the normal model before an installation environment change and the normal model after the installation environment change;
the inference unit calculates feature amounts of the normal model before the change in the installation environment and after the change in the installation environment, and compares the feature amounts of the normal model before the change in the installation environment and after the change in the installation environment;
the output unit outputs for display a comparison result of the feature quantity of the normal model before the change in the installation environment and after the change in the installation environment by the inference unit.
6. The factory automation device according to claim 2, claim 4, or claim 5. The factory automation equipment of claim 6;
a display that displays information about the power output from the output unit; and
A factory automation system comprising: A first step of extracting power from common mode noise generated between a frame ground of a housing and a signal ground of a signal circuit in a factory automation device;
a second step of generating a normal model for inferring whether or not an abnormality has occurred due to the common mode noise from time series data of information about power during stable operation of the factory automation equipment, among the extracted information about power;
A normal model generation method comprising: On the computer,
A first step of extracting power from common mode noise generated between a frame ground of a housing and a signal ground of a signal circuit in a factory automation device;
a second step of generating a normal model for inferring whether or not an abnormality has occurred due to the common mode noise from time series data of information about power during stable operation of the factory automation equipment, among the extracted information about power;
A program characterized by executing the following.