JP2022096212A - State estimation device, mounting system, and state estimation method - Google Patents

Abstract

To provide a state estimation device that can improve accuracy of detecting a head requiring maintenance.SOLUTION: A state estimation device 100 is a device that includes a head having a head main body part to which a nozzle configured to hold an object is attached, and manages a state of the head in equipment that holds the object with the head to perform predetermined operations, and the state estimation device comprises: a threshold creation unit 110 that acquires an equipment log of the head to which the nozzle is attached, and creates a first threshold on the basis of a dataset consisting of a plurality of characteristic values of the head included in the equipment log; and a state estimation unit 120 that acquires the characteristic values of the head to which the nozzle is attached and is in an operating state, and estimates whether the head is in a condition related to disorder on the basis of the acquired characteristic values of the head and the first threshold. The dataset consists of a plurality of characteristic values of the head exceeding a reference value.SELECTED DRAWING: Figure 1

Description

The present disclosure relates to state estimation devices, mounting systems, and state estimation methods.

Conventionally, equipment such as a component mounting system may deteriorate due to use or with the passage of time, so maintenance is required. For example, in Patent Document 1, the vacuum pressure of the vacuum pump at the head is measured by a sensor, and when the measured vacuum pressure reaches a threshold value when it is considered that the maintenance of the vacuum pump is necessary, the vacuum pump is maintained. The technology to determine that it is in time is disclosed.

Japanese Unexamined Patent Publication No. 2012-033522

However, the threshold value when it is considered that the maintenance of the head is considered to be necessary is often empirically obtained, and the threshold value is an accurate value with respect to the measured value of the sensor depending on the environment in which the equipment equipped with the head is installed. May not be. For this reason, there are cases where the head does not actually malfunction even if the measured value of the sensor reaches the threshold value, or the head actually malfunctions even though the measured value of the sensor does not reach the threshold value. There is a problem that the detection accuracy of the head that requires maintenance is not sufficient because it may be seen.

Therefore, the present disclosure provides a state estimation device and the like that can improve the detection accuracy of a head that requires maintenance.

The state estimation device according to one aspect of the present disclosure includes a head having a head main body portion to which a nozzle configured to hold the object is mounted, and the head holds the object to perform a predetermined operation. It is a state estimation device that manages the state of the head in the equipment to be performed, and is a data set consisting of a plurality of characteristic values of the head included in the equipment log by acquiring the equipment log of the head to which the nozzle is mounted. A threshold generation unit that generates a first threshold value based on the above, and a characteristic value of the head that is in an operating state with the nozzle mounted are acquired, and based on the acquired characteristic value of the head and the first threshold value. The data set comprises a state estimation unit for estimating whether or not the head is in a state related to a malfunction, and the data set comprises a plurality of characteristic values of the head exceeding a reference value.

It should be noted that these comprehensive or specific aspects may be realized in a system, device, method, recording medium, or computer program, and any combination of the system, device, method, recording medium, and computer program. It may be realized by.

According to the state estimation device and the like according to the present disclosure, it is possible to improve the detection accuracy of the head that requires maintenance.

It is a block diagram which shows an example of the state estimation apparatus which concerns on embodiment. It is a block diagram which shows an example of the mounting system which concerns on embodiment. It is a block diagram which shows an example of the component mounting apparatus which concerns on embodiment. It is a schematic diagram which shows the structure at the time of vacuum of the head main body part which concerns on embodiment. It is a schematic diagram which shows the structure at the time of blowing of the head main body part which concerns on embodiment. It is a figure for demonstrating the operation of the state estimation apparatus which concerns on embodiment. It is a figure for demonstrating the generation method of the online flow rate decrease determination threshold value. It is a figure for demonstrating the generation method of the offline flow rate drop determination threshold value. It is a figure for demonstrating the estimation method of the upset factor of a head. It is a graph which shows an example of the waveform data of the flow rate in an air path. It is a sequence diagram which shows an example of the operation of the equipment maintenance system and the person in charge of maintenance which concerns on embodiment. It is a flowchart which shows an example of the state estimation method which concerns on other embodiment.

The state estimation device of the present disclosure includes a head having a head main body portion to which a nozzle configured to hold the object is mounted, and the head holds the object by the head to perform a predetermined operation. It is a state estimation device that manages the state of the head, acquires the equipment log of the head to which the nozzle is mounted, and is the first based on a data set consisting of a plurality of characteristic values of the head included in the equipment log. A threshold generation unit that generates one threshold and a characteristic value of the head that is in an operating state with the nozzle attached are acquired, and the head is based on the acquired characteristic value of the head and the first threshold. The data set comprises a state estimation unit for estimating whether or not the state is related to a malfunction, and the data set comprises a plurality of characteristic values of the head exceeding a reference value.

For example, whether or not the equipment log of the head when the equipment performs a predetermined work (for example, production) using the head equipped with the nozzle is accumulated in the equipment and the head is in a state related to the malfunction. A first threshold for estimation is generated based on a dataset consisting of a plurality of characteristic values (eg, a large number of characteristic values) of the head included in such equipment logs. Then, the characteristic value of the head in the equipment currently in the operating state by performing a predetermined work is acquired, and the acquired characteristic value is compared with the first threshold value, so that the head is currently in a state related to the malfunction. It is estimated whether or not. In the present disclosure, the threshold value for estimating whether or not the head is in a state related to the malfunction is not empirically determined, but is objectively generated based on the equipment log, and therefore maintenance is required. The detection accuracy of the head can be improved. Further, the plurality of characteristic values of the head included in the equipment log are data that can be acquired without mechanically modifying the equipment or the like. Therefore, it is possible to improve the detection accuracy of the head that requires maintenance without mechanically modifying the equipment or the like.

Further, the threshold generation unit allocates the amount of state change of the head calculated by comparing a plurality of characteristic values of the head included in the equipment log with the reference value to a plurality of clusters as learning data. The first threshold value may be generated based on the characteristic value of the head in one cluster having the smallest amount of state change among the plurality of clusters.

For example, when the head is not able to hold the object correctly, the amount of state change tends to be large, and when the head is in a state related to the malfunction, the amount of state change tends to be small. That is, the characteristic value of the head in the cluster with a large amount of state change is likely to be the characteristic value when there is no malfunction in the head, and the characteristic value of the head in the cluster with a small amount of state change is related to the malfunction of the head. It is highly possible that it is a characteristic value when it is in a state. Therefore, by generating the first threshold value based on the characteristic value of the head in the cluster having the smallest amount of state change, it is possible to further improve the detection accuracy of the head that requires maintenance.

Further, the threshold value generation unit further acquires an inspection log of the head main body portion to which the nozzle is not mounted, and sets a second threshold value based on a plurality of characteristic values of the head main body portion included in the inspection log. When the head is estimated to be in a state related to the malfunction, the state estimation unit acquires the inspection result of the head main body portion to which the nozzle is not mounted, and obtains the inspection result and the second. Based on the threshold value, it may be estimated whether the nozzle is in a malfunctioning state or the head main body portion is in a malfunctioning state. For example, the plurality of characteristic values of the head main body portion are data measured by removing the nozzle from the head whose characteristic value of the head to which the nozzle is mounted exceeds the reference value, and is the threshold value. The generation unit performs the first threshold value by linear regression using a plurality of characteristic values of the head to which the nozzle is mounted and a plurality of characteristic values of the head main body portion corresponding to the plurality of characteristic values as learning data. The second threshold value may be generated based on the above.

For example, the inspection log of the head main body when the nozzle is removed from the head is accumulated in the equipment, and the nozzle is in a bad state when it is estimated that the head is in a bad state. A second threshold for estimating whether there is or the head body is in a malfunctioning state is based on a plurality of characteristic values (for example, a large number of characteristic values) of the head body included in such an inspection log. Is generated. Then, the inspection result of the head main body (specifically, the latest characteristic value of the head main body) in which the nozzle is removed from the head estimated to be in a state related to the malfunction is acquired, and the acquired inspection result is the first. By comparing with the two threshold values, it is estimated whether the nozzle is in a malfunctioning state or the head main body is in a malfunctioning state. In this way, it is possible to estimate the cause of the head malfunction that requires maintenance.

Further, the state estimation unit may presume that the head main body is in a malfunctioning state when the inspection result of the head main body exceeds the second threshold value. Further, the state estimation unit may estimate that the nozzle is in a malfunctioning state when the inspection result of the head main body portion does not exceed the second threshold value.

In this way, when the inspection result of the head main body exceeds the second threshold value, it can be estimated that the head main body is in a malfunctioning state, and the inspection result of the head main body exceeds the second threshold. If not, it can be estimated that the nozzle is in a bad condition.

Further, when the state estimation unit estimates that the head main body is in a malfunctioning state, the state estimation unit performs a correlation analysis using the factor estimation value calculated from the inspection log and the malfunctioning factor of the head main body as learning data. The malfunction factor of the head main body may be estimated based on the factor estimation value calculated from the inspection result of the head main body using the generated malfunction factor estimation table.

For example, a factor estimation value is calculated for each of a plurality of characteristic values of the head main body included in the inspection log, and a malfunction factor estimation table is generated in which the malfunction factor is associated with each calculated factor characteristic value. Then, the factor estimation value is calculated from the latest inspection result of the head main body part in which the nozzle is removed from the head estimated to be in a state related to the malfunction, and the factor estimation value is collated with the malfunction factor estimation table. , It is possible to estimate the cause of the malfunction of the head body.

Further, the factor estimation value calculated from the inspection result of the head main body may be calculated based on the statistical value of the inspection log determined that the head main body is in a normal state.

In this way, the latest inspection of the head body with the nozzle removed from the head estimated to be in a state related to the malfunction based on the statistical values of the inspection log determined that the head body is in a normal state. Factor estimates for the results can be calculated. For example, a unit space indicating the range of the normal state can be generated from the inspection log determined that the head main body is in the normal state, and the Mahalanobis distance can be used as a factor estimation value.

Further, the characteristic value may be a measurement result of a flow rate or a pressure in the air path of the head main body when air is supplied from the air source of the head main body at a positive pressure or a negative pressure.

According to this, using the measurement result of the flow rate or the pressure in the air path of the head main body, it is estimated whether or not the head is in a state related to the malfunction, and whether or not the head is in the malfunction state. It is possible to estimate the cause of the malfunction of the head or the cause of the malfunction of the head body.

Further, an output unit that outputs information based on the estimation result of the state estimation unit may be further provided.

According to this, it is possible to take measures according to the estimation result of the state estimation unit.

Further, the information may include information indicating that the head is in a malfunction-related state or the head is in a malfunction state.

According to this, it is possible to make the person in charge of maintenance of the equipment or the like aware that the head is in a state related to the malfunction or the head is in the malfunction state.

Further, the information may include maintenance contents for the head estimated to be in a malfunctioning state.

According to this, it is possible to perform maintenance according to the maintenance content for the head estimated to be in a malfunctioning state.

Further, the information may be output to the equipment maintenance planning unit that manages the maintenance plan for maintaining the equipment.

According to this, the equipment maintenance planning unit can manage the maintenance plan of the equipment according to the output information.

The mounting system of the present disclosure manages a head having a head main body to which a nozzle configured to hold an object is mounted, a measuring unit for measuring characteristic values of the head, and a state of the head. It is equipped with a state estimation device.

According to this, it is possible to provide a mounting system capable of improving the detection accuracy of a head that requires maintenance.

The state estimation method of the present disclosure comprises a head having a head main body portion to which a nozzle configured to hold an object is mounted, and the head holds the object by the head to perform a predetermined operation. It is a state estimation method that manages the state of the head, and obtains the equipment log of the head in which the nozzle is mounted on the head main body, and data including a plurality of characteristic values of the head included in the equipment log. The threshold generation step of generating the first threshold value based on the set and the characteristic value of the head in which the nozzle is attached to the head main body and are in an operating state are acquired, and the characteristic value of the head and the first one are acquired. The dataset comprises a plurality of characteristic values of the head that exceed a reference value, including a state estimation step that estimates whether or not the head is in a state associated with the malfunction based on a threshold.

According to this, it is possible to provide a state estimation method that can improve the detection accuracy of a head that requires maintenance.

It should be noted that all of the embodiments described below show comprehensive or specific examples. The numerical values, shapes, materials, components, arrangement positions and connection forms of the components, steps, the order of steps, and the like shown in the following embodiments are examples, and are not intended to limit the present disclosure.

(Embodiment)
Hereinafter, embodiments will be described with reference to FIGS. 1 to 9.

First, the state estimation device 100 according to the embodiment will be described.

FIG. 1 is a configuration diagram showing an example of a state estimation device 100 according to an embodiment.

The state estimation device 100 is a device that manages the state of the head in the equipment. The equipment comprises a head configured to hold the object (specifically, a head having a head body configured to hold the object), and the head holds the object to a predetermined position. Equipment for performing work, for example, a component mounting device. For example, a nozzle is attached to the head (head body), and the nozzle holds an object. The component mounting device holds the component as an object and performs mounting work (that is, production of the mounting board or the like) on the board or the like held as a predetermined operation. For example, the state estimation device 100 is a computer (server or the like) installed in a facility different from the facility in which the equipment is installed, but may be installed in the facility in which the equipment is installed. Further, the state estimation device 100 may be a computer provided in one housing, or may be divided into two or more housings and realized by two or more computers.

The state estimation device 100 includes a threshold value generation unit 110, a state estimation unit 120, and an output unit 130. The state estimation device 100 is realized by a computer including a processor, a memory, and the like. The threshold value generation unit 110, the state estimation unit 120, and the output unit 130 are realized by operating the processor according to a program stored in the memory.

The threshold value generation unit 110 acquires the equipment log of the head to which the nozzle is mounted, and generates a first threshold value based on a data set including a plurality of characteristic values of the head included in the equipment log. The data set consists of multiple characteristic values of the head that exceed the reference value. The reference value will be described later. In other words, the equipment log of the head on which the nozzle is mounted is a mounting in which the equipment holds the object by the head and mounts the parts held by the nozzle mounted on the head on a board or the like. This is data that records the operation results (errors, events, etc.) of the head when performing work). The characteristic value of the head is the measurement result of the flow rate or the pressure in the air path of the head. Specifically, the characteristic value of the head is a measurement result of a flow rate or a pressure in an air path of a nozzle and a head body when air is supplied from an air source at a positive pressure or a negative pressure. The air source is provided outside the component mounting device or the component mounting device. Here, the characteristic value of the head will be described as a flow rate value. Since the fact that the equipment is performing a predetermined work, that is, that the mounting work is in progress, can be called "online", the characteristic value of the head included in the equipment log is called an online flow rate value. Since the threshold value generation unit 110 acquires old online flow rate values in time series from the equipment log, the characteristic value of the head input to the threshold value generation unit 110 is referred to as a past online flow rate value.

Further, the threshold value generation unit 110 acquires an inspection log of the head main body portion to which the nozzle is not mounted, and generates a second threshold value based on a plurality of characteristic values of the head main body portion included in the inspection log. The plurality of characteristic values of the head main body are data measured by removing the nozzle from the head in which the characteristic value of the head to which the nozzle is mounted exceeds the reference value. When the characteristic value exceeds the reference value, the characteristic value exceeds the lower limit value or the upper limit value in the normal range. When the reference value is the lower limit value, the characteristic value exceeding the reference value means that the characteristic value is below the reference value. When the reference value is the upper limit value, the characteristic value exceeding the reference value means that the characteristic value exceeds the reference value. In other words, the inspection log of the head body to which the nozzle is not mounted is the head body when the equipment does not hold the object by the head and does not perform the predetermined work (specifically, the mounting work is not performed). It is a log of the data of the department. Here, the data is the inspection result when the inspection is performed on the head main body portion. Since the fact that the equipment does not perform the predetermined work, that is, the fact that the mounting work is not performed, can be called "offline", the characteristic value of the head main body portion included in the inspection log is called an offline flow rate value. Since the threshold value generation unit 110 acquires old offline flow rate values in time series from the inspection log, the characteristic value of the head main body unit input to the threshold value generation unit 110 is referred to as a past offline flow rate value.

Further, the threshold value generation unit 110 generates a malfunction factor estimation table in which the factor estimation value calculated from the inspection log and the malfunction factor based on the maintenance record of the head main body are associated with each other.

The details of the threshold value generation unit 110 will be described later.

The state estimation unit 120 has a first estimation unit 121, a second estimation unit 122, and a factor estimation unit 123, and the head state is estimated by the first estimation unit 121, the second estimation unit 122, and the factor estimation unit 123.

The first estimation unit 121 acquires the characteristic value of the head in which the nozzle is mounted and is in an operating state, and the head is based on the acquired characteristic value of the head and the first threshold value generated by the threshold value generation unit 110. Estimate whether or not the condition is related to the malfunction. A condition related to a malfunction is a condition in which there may be a sign of abnormality before an abnormal condition such as stopping production by equipment is reached. The first estimation unit 121 acquires the characteristic value of the head to which the nozzle newly output from the equipment performing the predetermined work is mounted. The characteristic value of the head acquired by the first estimation unit 121 is referred to as the latest online flow rate value with respect to the past online flow rate value. The details of the first estimation unit 121 will be described later.

The second estimation unit 122 acquires the inspection result of the head main body portion to which the nozzle is not mounted when it is estimated that the head is in a state related to the malfunction, and based on the inspection result and the second threshold value, the second estimation unit 122 obtains the inspection result. It is estimated whether the nozzle is in a malfunctioning state or the head body is in a malfunctioning state. The second estimation unit 122 is a head when the production of the equipment is stopped, the nozzle is removed from the head main body, and the equipment is not performing a predetermined operation after the head is estimated to be in a state related to the malfunction. Acquire the inspection result of the flow rate of the air path of the main body. The inspection result of the head main body portion acquired by the second estimation unit 122 is referred to as the latest offline flow rate value with respect to the past offline flow rate value. The details of the second estimation unit 122 will be described later.

The factor estimation unit 123 estimates the malfunction factor of the head main body portion based on the factor estimation value calculated from the inspection result of the head main body portion and the malfunction factor estimation table. The details of the factor estimation unit 123 will be described later.

The output unit 130 outputs information based on the estimation result of the state estimation unit 120.

For example, the output unit 130 outputs information to the maintenance person, and the information output from the output unit 130 includes a maintenance instruction to the maintenance person.

Further, the information output from the output unit 130 includes the maintenance content for the head estimated to be in a malfunctioning state. The information including the maintenance content output from the output unit 130 is output to, for example, the equipment maintenance planning unit that manages the equipment maintenance plan. For example, the information may be output to a maintenance device that maintains the equipment. For example, the maintenance content for the head includes the serial number of the head to be maintained, the installation location of the head to be maintained, the content of the maintenance work, the person in charge, the maintenance deadline, and the like.

Further, the information output from the output unit 130 may include information indicating that the head is in a state related to the malfunction or the head is in the malfunction state. The information including these information output from the output unit 130 is shown, for example, in a form recognizable by the maintenance person in the equipment connected to the state estimation device 100 or the state estimation device 100. For example, based on the estimation result by the first estimation unit 121, characters such as "the head may be out of order" or "the head needs to be inspected" may be displayed on the display or output from the speaker or the like. Alternatively, a buzzer or a lamp that is known in advance to indicate these contents may sound. Further, for example, based on the estimation result by the second estimation unit 122, characters such as "head body unit is malfunctioning" or "nozzle malfunction" are displayed on a display or the like, output from a speaker or the like, or A buzzer or a lamp, which is known in advance to indicate these contents, sounds or lights up. Further, for example, based on the estimation result by the factor estimation unit 123, characters such as "filter clogging has occurred in the head body" or "air hose has been torn in the head body" are displayed on the display or the like. Or, it is output from a speaker or the like, or a buzzer or a lamp that is known in advance to indicate these contents sounds or lights up. These displays, speakers, buzzers, etc. may be installed in the facility where the equipment is installed, and the maintenance personnel of the facility may be notified of these alerts directly.

By outputting the information based on the estimation result of the state estimation unit 120, it is possible to take measures according to the estimation result of the state estimation unit. For example, it is possible to make the maintenance personnel of the equipment recognize that the head is in a state related to the malfunction, or that the head is in the malfunction state, and for example, the head is in a malfunction state or a state related to the malfunction. It is possible to perform maintenance according to the maintenance content for the head estimated to be.

For example, the information output from the output unit 130 may include information about the head to be inspected based on the estimation result of the first estimation unit 121. The output unit 130 may output a maintenance instruction instructing the inspection of the head to be inspected based on the maintenance plan assigned to the inspection outside the time when the equipment performs a predetermined work. The state estimation device 100 can notify the maintenance person that the head needs to be inspected. The time when the equipment does a predetermined work is, for example, the time when the equipment is not in production. Since the head inspection needs to be done with the nozzle removed, the head inspection is done at a time when the equipment is not in production. The time when the equipment is not in production may be the timing of switching the varieties to be produced. As a result, it is possible to inspect the head to be inspected outside the time when the equipment performs a predetermined work (for example, production). The head inspection may also be performed outside the equipment. In this case, it is possible to inspect the head to be inspected during the time when the equipment is in production.

As described above, the equipment may be a component mounting device, and the state estimation device 100 may configure a mounting system together with the component mounting device. Here, the mounting system 2 according to the embodiment will be described.

FIG. 2A is a configuration diagram showing an example of the mounting system 2 according to the embodiment.

The mounting system 2 is a system for mounting components on a board or the like, and includes a component mounting device 300 having a head 310 and a control unit 330, and a state estimation device 100.

The head 310 has a head main body 311 and a sensor 320, and a nozzle 312 can be attached to the head main body 311. The sensor 320 is an example of a measuring unit that measures the characteristic value of the head 310 (head main body unit 311). The sensor 320 is a flow rate sensor that measures the flow rate in the air path of the head main body 311. The sensor 320 has an online flow rate when the component mounting device 300 is performing a predetermined work by holding the object by the head 310 to which the nozzle 312 is mounted, and the component mounting device in which the nozzle 312 is not mounted to the head 310. The offline flow rate is measured when the 300 is not performing a predetermined operation. Further, the head 310 may have a vacuum sensor 340 for measuring the vacuum pressure in the air path of the head main body 311 (see FIG. 2B).

The control unit 330 controls the head 310. Details of the control of the head 310 will be described later. Further, the control unit 330 acquires the online flow rate value and the offline flow rate value measured by the sensor 320. The control unit 330 outputs the online flow rate value and the offline flow rate value to the state estimation device 100 when the state estimation device 100 generates the first threshold value, the second threshold value, and the malfunction factor estimation table. Further, the control unit 330 outputs the latest online flow rate value and the latest offline flow rate value to the state estimation device 100 when the state estimation device 100 estimates the state of the head 310.

Here, a specific example of the component mounting device 300 will be described with reference to FIG. 2B.

FIG. 2B is a configuration diagram showing an example of the component mounting device 300 according to the embodiment. In FIG. 2B, the X direction of the substrate transport direction (the direction perpendicular to the paper surface in FIG. 2B) and the Y direction orthogonal to the substrate transport direction (the left-right direction in FIG. 2B) are shown as biaxial directions orthogonal to each other in the horizontal plane. Further, the Z direction (vertical direction in FIG. 2B) is shown as a height direction orthogonal to the horizontal plane.

The component mounting device 300 has a function of mounting the component D on the substrate B. The substrate transport mechanism 12 provided on the upper surface of the base 11 transports the substrate B in the X direction, positions it, and holds it. The head moving mechanism 13 moves the head 310 mounted via the plate 13a in the X direction and the Y direction. A nozzle 312 is attached to the lower end of the head 310.

A plurality of tape feeders 16 are attached side by side in the X direction on the upper portion of the carriage 17 coupled to the base 11 on the side of the substrate transport mechanism 12. In the carriage 17, a carrier tape 18 for storing a component D supplied to the component mounting device 300 is wound and stored on a reel 19 and held. The carrier tape 18 inserted in the tape feeder 16 is pitch-fed at regular intervals by the tape feeding mechanism 16a built in the tape feeder 16. As a result, the component D stored in the carrier tape 18 is sequentially supplied to the component supply port 16b provided on the upper portion of the tape feeder 16.

In FIG. 2B, the component mounting device 300 includes a control unit 330 that controls a board transfer mechanism 12, a head moving mechanism 13, a head 310, and a tape feeder 16 to execute a component mounting operation. In the component mounting operation, the control unit 330 moves the head 310 above the tape feeder 16 by the head moving mechanism 13, and vacuum sucks and picks up the component D supplied by the tape feeder 16 to the component supply port 16b by the nozzle 312. (Arrow a). Next, the control unit 330 moves the head 310 holding the component D by the head moving mechanism 13 above the substrate B held by the substrate transport mechanism 12, and mounts the component D at a predetermined component mounting position Ba on the substrate B. (Arrow b).

In FIG. 2B, the head 310 includes a vacuum sensor 340 that measures the degree of vacuum when the nozzle 312 vacuum sucks the component D. From the measurement result of the degree of vacuum of the nozzle 312 during the component holding operation by the vacuum sensor 340, it is possible to detect the presence or absence of a vacuum error such as a suction error (suction error) or a malfunction of the head 310. For example, when the nozzle 312 normally adsorbs the component D, the degree of vacuum becomes smaller than the predetermined value, and when the nozzle 312 cannot hold the component D or adsorbs the component D in an abnormal posture, the degree of vacuum does not drop to the predetermined value. Therefore, the control unit 330 can detect a vacuum error by determining whether or not the degree of vacuum measured by the vacuum sensor 340 exceeds a predetermined value.

In FIG. 2B, a substrate recognition camera 20 whose optical axis direction is directed downward is attached to the plate 13a. The substrate recognition camera 20 moves in the X direction and the Y direction integrally with the head 310 by the head moving mechanism 13. The board recognition camera 20 moves above the tape feeder 16 and takes an image of the component D supplied to the supply position of the component supply port 16b.

The control unit 330 recognizes the image pickup result and calculates the amount of supply position deviation in which the actually supplied component D deviates from the expected normal supply position. Further, the control unit 330 corrects the suction position (stop position of the head 310) when the nozzle 312 picks up the component D or the supply position of the component D of the tape feeder 16 based on the calculated supply position deviation amount. .. Further, the control unit 330 recognizes the image pickup result as an image, and detects a supply error in which the component D cannot be recognized because the component D is not supplied to the component supply port 16b.

In FIG. 2B, a component recognition camera 21 with the optical axis direction facing upward is attached to the upper surface of the base 11 between the substrate transfer mechanism 12 and the tape feeder 16. The component recognition camera 21 captures an image of the lower surface of the component D (or the nozzle 312 that could not hold the component D) held by the nozzle 312 when the nozzle 312 that picks up the component D passes above.

The control unit 330 recognizes the image pickup result as an image, and an error occurs in which the posture of the component D held by the nozzle 312 is abnormal or the component D which should be held by the nozzle 312 cannot be recognized. Determine if you are not. Further, the control unit 330 recognizes the image pickup result and calculates the suction position deviation amount in which the component D actually sucked on the nozzle 312 is displaced from the expected normal suction position. When the component D is mounted on the component mounting position Ba on the substrate B, the control unit 330 executes mounting position correction and mounting posture correction based on the suction position deviation amount.

As described above, the component mounting device 300 includes a tape feeder 16, a head 310, and a nozzle 312. Then, the control unit 330 has detected suction error, supply error, recognition error occurrence status, calculated supply position deviation amount, correction amount of suction position of component D by nozzle 312, suction position deviation amount, mounting position correction amount, The wearing posture correction amount, etc. is transmitted to the management computer, etc. in association with the device status (normal, abnormal, etc.). Further, the control unit 330 transmits the measurement results of various sensors included in each mechanism of the component mounting device 300, such as the degree of vacuum by the vacuum sensor 340, to the state estimation device 100 in association with, for example, the device status (normal, abnormal, etc.). do.

Here, the configuration of the head main body portion 311 will be described with reference to FIGS. 3A and 3B.

FIG. 3A is a schematic view showing the configuration of the head main body portion 311 according to the embodiment at the time of vacuum.

FIG. 3B is a schematic view showing a configuration of the head main body portion 311 according to the embodiment at the time of blowing.

3A and 3B schematically show a nozzle 312 mounted on the head body 311 and an air source 319 for supplying air to the head body 311 in addition to the head body 311.

The head body 311 includes a nozzle holder 313, a blow valve 314, a vacuum valve 315, a common air path 316, a blow air path 317 and a vacuum air path 318.

A nozzle 312 is attached to the nozzle holder 313, and parts are attracted via the nozzle 312.

The air source 319 is a device that supplies air with a positive pressure or a negative pressure.

The blow valve 314 is a valve that controls the supply of air to the nozzle holder 313 when the air source 319 is supplying air with a positive pressure (that is, at the time of blowing).

The vacuum valve 315 is a valve that controls suction of air from the nozzle holder 313 when the air source 319 supplies air with a negative pressure (that is, at the time of vacuum).

The common air path 316 is from the nozzle holder 313 when the air from the air source 319 passes when the air source 319 is supplying air with a positive pressure and when the air source 319 is supplying air with a negative pressure. This is the air path through which the air passes. That is, the common air path 316 is a path through which air passes in common during blow and vacuum.

The blow air path 317 is an air path through which air from the air source 319 passes when the air source 319 is supplying air with a positive pressure.

The vacuum air path 318 is an air path through which air from the nozzle holder 313 passes when the air source 319 supplies air with a negative pressure.

The blow valve 314 and the vacuum valve 315 have a switching mechanism for switching the air path between blow time and vacuum time. For example, as shown in FIG. 3A, during vacuum, the switching mechanism in the blow valve 314 and vacuum valve 315 is such that the vacuum air source is connected to the nozzle holder 313 via the vacuum air path 318 and the common air path 316. Be controlled. For example, as shown in FIG. 3B, the switching mechanism in the blow valve 314 and the vacuum valve 315 is controlled so that the blow air source is connected to the nozzle holder 313 via the blow air path 317 and the common air path 316 during blow. To.

When the component is sucked by the nozzle 312 attached to the nozzle holder 313, the vacuum valve 315 is controlled to suck air from the nozzle 312 to the air source 319 via the common air path 316 and the vacuum air path 318. When the suction of the parts by the nozzle 312 attached to the nozzle holder 313 is released, the vacuum valve 315 is controlled, the path from the air source 319 to the vacuum air path 318 is closed, and the suction of the parts by the nozzle 312 is closed. It will be released. Even if the nozzle 312 is not mounted on the nozzle holder 313 and the parts cannot be sucked, air is sucked in order to acquire the inspection result (that is, the offline flow rate) of the head main body portion 311.

When supplying air, the blow valve 314 is controlled to supply air from the air source 319 to the nozzle 312 via the blow air path 317 and the common air path 316. Even when the nozzle 312 is not attached to the head main body 311, air is supplied in order to acquire the inspection result (that is, the offline flow rate) of the head main body 311.

The sensor 320 is a flow rate sensor, measures the flow rate in the air path (specifically, the common air path 316 and the blow air path 317) when air is supplied from the air source 319 to the nozzle holder 313, and air from the nozzle holder 313. The flow rate in the air path (specifically, the common air path 316 and the vacuum air path 318) when air is sucked into the source 319 is measured. Further, by controlling the air source 319, the blow valve 314 and the vacuum valve 315, the common air path 316, the blow air path 317 and the vacuum air path 318 can be evacuated.

Next, the operation of the state estimation device 100 will be described.

FIG. 4 is a diagram for explaining the operation of the state estimation device 100 according to the embodiment. As the phase in which the operation of the state estimation device 100 is performed, there are an operation of the threshold value generation unit 110 in the learning phase and an operation of the state estimation unit 120 in the estimation phase. Further, as the operation contents of the state estimation device 100, an online flow rate change detection block in which a first threshold value is generated by the threshold value generation unit 110 and the first estimation unit 121 estimates using the generated first threshold value, and a threshold value. A second threshold value is generated by the generation unit 110, a flow rate change factor unit estimation block for estimation by the second estimation unit 122 using the generated second threshold value, and a malfunction factor estimation table are generated by the threshold value generation unit 110. There is a head malfunction factor location estimation block in which the factor estimation unit 123 estimates using the generated malfunction factor estimation table.

First, the online flow rate change detection block will be described.

In the online flow rate change detection block in the learning phase, the threshold generation unit 110 acquires equipment logs from the component mounting device 300. The equipment log contains past online flow values. The past online flow rate value acquired by the online flow rate change detection block is the online flow rate value when the vacuum sensor 340 detects a vacuum error in the air path. The vacuum error threshold for determining whether or not the vacuum sensor 340 detects a vacuum error is, for example, a value empirically obtained. Depending on the environment in which the component mounting device 300 is installed, the vacuum error threshold is a vacuum. It may not be an accurate value with respect to the measured value of the sensor 340. Therefore, even if the measured value of the vacuum sensor 340 reaches the vacuum error threshold value, the head 310 may not actually malfunction, or the measured value of the vacuum sensor 340 does not reach the vacuum error threshold value. However, there is a problem that the detection accuracy of the head 310, which requires maintenance, is not sufficient because the head 310 may actually have a malfunction.

Therefore, the threshold value generation unit 110 uses the vacuum error threshold value to generate an online flow rate decrease determination threshold value that can estimate the malfunction of the head 310 with higher accuracy than the vacuum error threshold value. The online flow rate decrease determination threshold value is an example of the first threshold value, and the vacuum error threshold value is an example of a reference value to be compared with the online flow rate value.

For example, the threshold value generation unit 110 allocates the difference between the online flow rate value of the head 310 and the vacuum error threshold value calculated by comparing the past online flow rate value with the vacuum error threshold value to a plurality of clusters as learning data, and a plurality of clusters. The online flow rate decrease determination threshold value is generated based on the online flow rate value in one of the clusters having the smallest difference. The difference between the online flow rate value and the vacuum error threshold is an example of the amount of state change. Here, a method of generating the online flow rate decrease determination threshold value will be described with reference to FIG.

FIG. 5 is a diagram for explaining a method of generating an online flow rate decrease determination threshold value. In FIG. 5, the online flow rate value included in the equipment log is plotted with the horizontal axis representing the online flow rate value and the vertical axis representing the difference between the online flow rate value and the vacuum error threshold value.

The threshold generation unit 110 classifies the acquired past online flow rate values into classes. For example, the threshold value generation unit 110 classifies the difference between the online flow rate value and the vacuum error threshold value when a vacuum error is detected by applying the k-means method into two classes. For example, as shown in FIG. 5, it is assumed that the class A has a large difference between the online flow rate value and the vacuum error threshold value and the class B has a small difference between the online flow rate value and the vacuum error threshold value. If the head 310 does not properly adsorb the parts, the air paths (specifically, the common air path 316 and the vacuum air path 318) communicate with the outside, so that a large amount of air is sucked and the above difference tends to increase. It is in. Therefore, the class A is a class in which there is a high possibility that the head 310 does not properly adsorb parts, in other words, a class in which there is a high possibility that the head 310 itself has no malfunction. On the other hand, when the head 310 correctly adsorbs the component, the amount of air sucked is small, and the above difference tends to be small. Therefore, the class B is a class in which there is a high possibility that the head 310 has a malfunction. Therefore, the threshold value generation unit 110 generates an online flow rate decrease determination threshold value based on the online flow rate value in the class B in which the difference is the smallest. Specifically, the threshold generation unit 110 subtracts 1 from the minimum value of the online flow rate value in class B (point P: −70 in FIG. 5) to determine the online flow rate decrease determination threshold value (broken line th: in FIG. 5). -71). When generating the online flow rate decrease determination threshold value, the value to be subtracted from the minimum value of the online flow rate value in the class having the smallest difference is not limited to 1, and is appropriately selected.

In the online flow rate change detection block in the estimation phase, the first estimation unit 121 acquires the latest online flow rate value of the head 310 to which the nozzle 312 is mounted, and is based on the acquired online flow rate value and the online flow rate decrease determination threshold value. It is estimated whether or not the head 310 is in a state related to the malfunction.

The first estimation unit 121 is when the acquired latest online flow rate value exceeds the online flow rate decrease determination threshold value (specifically, the absolute value of the online flow rate value is smaller than the absolute value of the online flow rate decrease determination threshold value). Estimates that the head 310 is in a state related to the malfunction, and the output unit 130 outputs to that effect. For example, the maintenance person or the like is notified that the head 310 is presumed to be in a state related to the malfunction, and determines whether to perform the inspection of the head 310.

The online flow rate decrease determination threshold value is an estimated value for indirectly determining the presence or absence of an abnormality such as clogging of the flow rate system, which is derived by class classification using the past online flow rate value as learning data. In the online flow rate change detection block, since it is estimated whether or not the head 310 is in a state related to the malfunction using such an online flow rate decrease determination threshold value, the head 310 is associated with the malfunction using the vacuum error threshold value. It is possible to improve the detection accuracy of the head 310, which requires maintenance, as compared with the case where it is estimated whether or not the head 310 is in a state.

Then, upon receiving the information regarding the malfunction of the head output from the output unit 130, the maintenance person inspects the head 310. For example, when the component mounting device 300 is not performing a predetermined operation (that is, production is stopped, for example), the nozzle 312 is removed from the head body 311 and the head body 311 to which the nozzle 312 is not mounted is removed. Be inspected. Specifically, the latest offline flow rate of the head main body 311 is measured. The latest measured offline flow rate value is used for estimation by the second estimation unit 122 in the flow rate change factor unit estimation block.

Next, the flow rate change factor unit estimation block will be described.

In the flow rate change factor unit estimation block in the learning phase, the threshold generation unit 110 acquires an inspection log from the component mounting device 300. The inspection log contains past offline flow values. The inspection log acquired by the flow rate change factor unit estimation block includes the past offline flow rate values when the vacuum sensor 340 detects a vacuum error in the air path.

For example, the threshold value generation unit 110 acquires the inspection log of the head main body unit 311 to which the nozzle 312 is not mounted, and determines the offline flow rate decrease determination threshold value based on the past offline flow rate value of the head main body unit 311 included in the inspection log. Generate. The offline flow rate decrease determination threshold value is an example of the second threshold value. Here, a method of generating an offline flow rate decrease determination threshold value will be described with reference to FIG.

FIG. 6 is a diagram for explaining a method of generating an offline flow rate decrease determination threshold value. FIG. 6 plots the combination of the online flow rate value and the offline flow rate value when a vacuum error is detected, which is included in the equipment log, with the horizontal axis representing the online flow rate value and the vertical axis representing the offline flow rate value.

The threshold generation unit 110 calculates a regression line from each of the plotted points, as shown in FIG. Since there is a correlation between the online flow rate value acquired when a vacuum error occurs and the offline flow rate value obtained when the nozzle 312 is removed from the head 310 and the head body 311 is inspected at that time, the online flow rate There is also a correlation between the decrease determination threshold and the offline flow rate decrease determination threshold. Therefore, the threshold value generation unit 110 sets the offline flow rate value corresponding to the online flow rate reduction determination threshold value as the offline flow rate decrease determination threshold value based on the calculated regression line. Specifically, as shown in FIG. 6, the threshold generation unit 110 has an offline flow rate value −90 on the vertical axis corresponding to the online flow rate value −71 (that is, the online flow rate decrease determination threshold value) on the horizontal axis on the regression line. .5 is generated as an offline flow rate decrease determination threshold value.

As described above, the threshold generation unit 110 uses the learning data as the plurality of online flow rate values of the head 310 to which the nozzle 312 is mounted and the plurality of offline flow rate values of the head main body unit 311 corresponding to the plurality of online flow rate values. An offline flow rate decrease determination threshold is generated based on the online flow rate decrease determination threshold by the linear regression.

In the flow rate change factor unit estimation block in the estimation phase, the second estimation unit 122 is a head body to which the nozzle 312 is not mounted when the head 310 is estimated to be in a state related to the malfunction by the first estimation unit 121. The inspection result of the unit 311 (that is, the latest offline flow rate value) is acquired, and the nozzle 312 is in a malfunction state or the head main body portion 311 is malfunctioning based on the acquired inspection result and the offline flow rate decrease determination threshold value. Estimate whether it is in a good state.

The second estimation unit 122 heads when the inspection result of the head main body unit 311 exceeds the offline flow rate decrease determination threshold value (specifically, the absolute value of the offline flow rate value is smaller than the absolute value of the offline flow rate decrease determination threshold value). It is estimated that the main body 311 is in a bad state. This is because even if the nozzle 312 is removed, the inspection result of the head main body portion 311 exceeds the offline flow rate decrease determination threshold value, and it can be estimated that the nozzle 312 side is not in a malfunctioning state but the head main body portion 311 side is in a malfunctioning state. For example, when air leakage or air clogging occurs in the air path of the head main body 311, the inspection result of the head main body 311 exceeds the offline flow rate decrease determination threshold even if the nozzle 312 is removed, and the head main body 311 It can be estimated that the side is in a bad state. On the other hand, the second estimation unit 122 estimates that the nozzle 312 is in a malfunctioning state when the inspection result of the head main body unit 311 does not exceed the offline flow rate decrease determination threshold value. By removing the nozzle 312, it can be estimated that the inspection result of the head main body 311 does not exceed the offline flow rate decrease determination threshold value, and it is said that the head main body 311 side is not in a malfunction state but the nozzle 312 side is in a malfunction state. This is because it can be estimated. For example, the output unit 130 outputs that the nozzle 312 is in a malfunctioning state, and the person in charge of maintenance of the equipment or the like can maintain the nozzle 312.

The offline flow rate decrease determination threshold value is an estimated value for indirectly determining the presence or absence of an abnormality such as clogging of the flow rate system, which is derived by linear regression using the past offline flow rate values as learning data. In the flow rate change factor unit estimation block, it is estimated whether the nozzle 312 is in a malfunctioning state or the head main body 311 is in a malfunctioning state by using such an offline flow rate decrease determination threshold value. The accuracy of estimation can be improved.

Next, the head malfunction factor location estimation block will be described.

In the head malfunction factor location estimation block in the learning phase, the threshold generation unit 110 acquires an inspection log from the component mounting device 300. The inspection log contains waveform data of past offline flow rates. For example, in the past offline flow rate waveform data acquired by the head malfunction factor location estimation block, in addition to the offline flow rate waveform data when the vacuum sensor 340 detects a vacuum error in the air path, the vacuum sensor 340 includes the air path. The waveform data of the offline flow rate at the normal time when the vacuum error of is not detected is included. The threshold value generation unit 110 generates a malfunction factor estimation table by correlation analysis using the factor estimation value calculated from the inspection log and the malfunction factor of the head main body portion 311 as learning data. The factor estimation value calculated from the inspection log and the malfunction factor estimation table will be described later.

In the head malfunction factor location estimation block in the estimation phase, when the factor estimation unit 123 estimates that the head body unit 311 is in a malfunction state by the second estimation unit 122, the factor estimation unit 123 uses the malfunction factor estimation table to use the head body unit. The cause of the malfunction of the head main body 311 is estimated based on the factor estimation value calculated from the inspection result of 311. Here, a method of estimating the cause of malfunction of the head 310 will be described with reference to FIG. 7.

FIG. 7 is a diagram for explaining a method of estimating the cause of malfunction of the head 310. The upper side of FIG. 7 shows a method of generating the average vector for factor estimation value calculation and the covariance matrix for factor estimation value calculation from the inspection log in the threshold generation unit 110, and the lower side of FIG. 7 shows a method of generating the covariance matrix for factor estimation value calculation. The factor estimation unit 123 shows a method of estimating a malfunction factor of the head main body unit 311 by using a factor estimation value calculated based on an average vector for factor estimation value calculation and a covariance matrix for factor estimation value calculation. ..

As shown in the upper part of FIG. 7, the threshold generation unit 110 acquires the waveform data set of the past offline flow rate at the normal time when the vacuum sensor 340 does not detect the vacuum error of the air path from the inspection log. Here, a specific example of the waveform data will be described with reference to FIG.

FIG. 8 is a graph showing an example of waveform data of the flow rate in the air path. FIG. 8 shows time-series data of flow rate values sampled at, for example, 256 points when the horizontal axis is time and the vertical axis is flow rate. FIG. 8 shows, as an example, waveform data of the flow rate in the air path when the vacuum valve 315 is repeatedly turned on and off during vacuum. Similar waveform data can be obtained during blow. In the present embodiment, when the waveform data of the past offline flow rate is acquired, it is acquired as a waveform data set defined by an array as shown in the upper part of FIG. 7.

Since the waveform data of the flow rate measured by the sensor 320 has a waveform shape as shown in FIG. 8, it is difficult to compare the waveform data with each other. Therefore, as shown on the upper side of FIG. 7, the threshold generation unit 110 makes it possible to compare the characteristics of each of the waveform data of the plurality of past offline flow rates with the characteristics of the waveform data of the latest offline flow rate. Each of the waveform data sets of the past offline flow rate is converted into a principal component vector by principal component analysis. Then, the threshold value generation unit 110 generates an average vector for factor estimation value calculation and a covariance matrix for factor estimation value calculation from the converted principal component vector.

Further, the threshold value generation unit 110 calculates the Mahalanobis distance of each of the principal component vectors converted from the waveform data set of the past offline flow rate as the factor estimation value. For example, it is possible to grasp from the maintenance record what kind of malfunction factor the head main body 311 is obtained from each of the principal component vectors, in other words, each of the waveform data sets. Therefore, it is possible to associate the malfunction factor based on the maintenance record with the factor estimation value of the principal component vector corresponding to each of the waveform data sets of the past offline flow rate. In this way, the threshold value generation unit 110 generates a malfunction factor estimation table in which the factor estimation value calculated from the inspection log and the malfunction factor based on the maintenance record of the head main body unit 311 are associated with each other.

As shown in the lower part of FIG. 7, the factor estimation unit 123 acquires the latest waveform data of the offline flow rate and converts it into a principal component vector by principal component analysis. Then, the factor estimation unit 123 calculates the Mahalanobis distance of the converted principal component vector. The factor estimation unit 123 uses the calculated Mahalanobis distance as a factor estimation value as an index for estimating the cause of the malfunction of the head body unit 311. In this way, the factor estimation value calculated from the inspection result of the head main body 311 is the statistical value of the inspection log determined that the head main body 311 is in a normal state (specifically, from the statistical value of the inspection log). Calculated based on the generated mean vector and variance matrix). The factor estimation unit 123 estimates the malfunction factor of the head main body portion 311 by collating the factor estimation value corresponding to the calculated latest offline flow rate waveform data with the malfunction factor estimation table. For example, when the factor estimation value equivalent to the calculated factor estimation value is in the malfunction factor estimation table, the factor estimation unit 123 heads the malfunction factor associated with the factor estimation value in the malfunction factor estimation table. It is presumed to be the cause of the malfunction of the part 311. For example, the output unit 130 outputs the cause of the malfunction of the head main body 311, and the person in charge of maintenance of the equipment can perform maintenance according to the cause of the malfunction of the head main body 311.

If the calculated factor estimation value is within the unit space (range of non-defective products), the head main body unit 311 may not be in a malfunction state, but the head 310 is malfunctioning in the first estimation unit 121. It is presumed that the head main body portion 311 is in a malfunctioning state in the second estimation unit 122, and there is a possibility that the head main body portion 311 is in a malfunctioning state. Therefore, the factor estimation unit 123 may estimate that the head main body unit 311 may be in a malfunctioning state when the calculated factor estimation value is in the unit space. For example, the output unit 130 outputs that the head main body unit 311 may be in a malfunction state, and the person in charge of maintenance of the equipment or the like can perform maintenance of the head main body unit 311.

The malfunction factor table is a correlation table generated by the correlation analysis of the factor estimation value obtained from the inspection log and the malfunction factor described in the maintenance record. In the head malfunction factor location estimation block, since the malfunction factor of the head main body 311 is estimated using such a malfunction factor estimation table, the accuracy of the estimation can be improved.

Next, the operation between the state estimation device 100 and the person in charge of maintenance of the equipment will be described with reference to FIG.

FIG. 9 is a sequence diagram showing an example of the operation of the state estimation device 100 and the maintenance person according to the embodiment.

First, the state estimation device 100 performs learning based on past data in advance (step S101). That is, the state estimation device 100 generates in advance an online flow rate decrease determination threshold value, an offline flow rate decrease determination threshold value, and a malfunction factor estimation table.

The state estimation device 100 periodically receives the latest online flow rate value from the component mounting device 300 (step S102), and performs processing in the online flow rate change detection block shown in FIG. 4 (step S103).

When the latest received online flow rate value is normal (that is, when the latest received online flow rate value does not exceed the online flow rate decrease determination threshold value), the state estimation device 100 continues the processing in steps S102 and S103. I do.

When the latest received online flow rate value is abnormal (that is, when the latest received online flow rate value exceeds the online flow rate decrease determination threshold value), the state estimation device 100 issues a flow rate change detection alert (step). S104). For example, in this case, the state estimation device 100 outputs information for carrying out the inspection of the head 310, specifically, an instruction for carrying out the inspection of the head 310. The timing at which the instruction to perform the inspection of the head 310 is transmitted can be appropriately determined.

The maintenance person receives the instruction to carry out the inspection of the head 310 (step S105), and carries out the inspection of the head 310 (step S106). For example, the maintenance person removes the nozzle 312 from the head main body 311 and inspects the head main body 311 (measures the latest offline flow rate).

The state estimation device 100 receives the latest offline flow rate value from the component mounting device 300 (step S107), and performs processing in the flow rate change factor unit estimation block shown in FIG. 4 (step S108).

When the latest received offline flow rate value is normal (that is, when the latest received online flow rate value does not exceed the offline flow rate decrease determination threshold value), the state estimation device 100 determines that the nozzle 312 is in a malfunctioning state. It estimates and notifies the nozzle maintenance instruction information which is the information for instructing the maintenance of the nozzle 312 (step S109).

The maintenance person receives the maintenance instruction of the nozzle 312 (step S110) and carries out the maintenance of the nozzle 312 (step S111). After the maintenance of the nozzle 312 is completed, the maintenance person in charge registers the completion of the maintenance of the nozzle 312 (step S112).

On the other hand, in the state estimation device 100, when the latest received offline flow rate value is abnormal (that is, when the latest received online flow rate value exceeds the offline flow rate decrease determination threshold value), the head main body 311 is malfunctioning. It is estimated that the state is in the above state, and processing is performed in the head malfunction factor location estimation block shown in FIG. 4 (step S113). The state estimation device 100 estimates the cause of the malfunction of the head main body 311 and notifies the head maintenance instruction information which is the information for instructing the maintenance of the head main body 311 (step S114). For example, the head maintenance instruction information includes a maintenance instruction of the head main body portion 311, specifically, a malfunction factor of the head main body portion 311.

The maintenance person receives the maintenance instruction of the head main body 311 (step S115) and carries out the maintenance of the head main body 311 (step S116). Since the maintenance instruction of the head main body 311 includes the cause of the malfunction of the head main body 3, the person in charge of maintenance can efficiently maintain the head main body 311 by performing the maintenance centering on the location of the cause of the malfunction. It can be carried out. After the maintenance of the head main body 311 is completed, the maintenance person registers the completion of the maintenance of the head main body 311 (step S117).

As described above, for example, the equipment log of the head 310 when the component mounting device 300 performs a predetermined work (for example, production) using the head 310 equipped with the nozzle 312 is accumulated in the component mounting device 300. An online flow rate reduction determination threshold for estimating whether or not the head 310 is in a state related to the malfunction is generated based on the online flow rate value (for example, a large amount of online flow rate value) included in such an equipment log. Will be done. Then, the online flow rate value in the component mounting device 300 that is currently in operation by performing a predetermined operation is acquired, and the acquired online flow rate value is compared with the online flow rate decrease determination threshold value, so that the head 310 is currently malfunctioning. It is estimated whether or not the condition is related to. In the present disclosure, the threshold value for estimating whether or not the head 310 is in a state related to the malfunction is not empirically determined, but is objectively generated based on the equipment log, and therefore maintenance is required. The detection accuracy of the head 310 can be improved. Further, the online flow rate value included in the equipment log is data that can be acquired without mechanically modifying the component mounting device 300 or the like. Therefore, it is possible to improve the detection accuracy of the head 310, which requires maintenance, without mechanically modifying the component mounting device 300 or the like.

Further, for example, the inspection log of the head main body 311 when the head main body 311 from which the nozzle 312 is removed from the head 310 is inspected is accumulated in the component mounting device 300, and it is estimated that the head 310 is in a malfunctioning state. An offline flow rate reduction determination threshold value for estimating whether the nozzle 312 is in a malfunctioning state or the head main body 311 is in a malfunctioning state when the nozzle is used is an offline flow rate value included in such an inspection log. Generated based on (eg, large offline flow values). Then, the inspection result (specifically, the latest offline flow rate value) of the head main body 311 from which the nozzle 312 is removed from the head 310 estimated to be in a state related to the malfunction is acquired, and the acquired inspection result is offline. By comparing with the flow rate decrease determination threshold value, it is estimated whether the nozzle 312 is in a malfunctioning state or the head main body portion 311 is in a malfunctioning state. In this way, it is possible to estimate the cause of malfunction of the head 310 that requires maintenance.

Further, for example, a factor estimation value is calculated for each of the waveform data of the offline flow rate included in the inspection log, and a malfunction factor estimation table is generated in which the malfunction factor is associated with each calculated factor characteristic value. Then, a factor estimation value is calculated from the latest offline flow rate waveform data of the head main body 311 from which the nozzle 312 is removed from the head 310 estimated to be in a state related to the malfunction, and the factor estimation value is the malfunction factor estimation table. By collating with, it is possible to estimate the cause of the malfunction of the head main body portion 311.

(Other embodiments)
Although the state estimation device 100 and the mounting system 2 of the present disclosure have been described above based on the embodiments, the present disclosure is not limited to the above-described embodiments. As long as it does not deviate from the gist of the present disclosure, various modifications that can be conceived by those skilled in the art are applied to the present embodiment, and a form constructed by combining components in different embodiments is also included in the scope of the present disclosure. Will be.

For example, in the above embodiment, the online flow rate value and the offline flow rate value have been described as the measured values of the sensor 320 when the vacuum sensor 340 detects a vacuum error, but the present invention is not limited to this. For example, the sensor 320 may detect an error with a value set regardless of the equipment environment, and the online flow rate value and the offline flow rate value are the measured values of the sensor 320 when the sensor 320 detects such an error. May be.

For example, in the above embodiment, the state estimation unit 120 has described the example of having the factor estimation unit 123, but the state estimation unit 120 may not have the factor estimation unit 123. In this case, the output unit 130 does not have to output information according to the estimation result of the factor estimation unit 123, and only outputs the information according to the estimation result of the first estimation unit 121 or the estimation result of the second estimation unit 122. It may be output.

For example, in the above embodiment, the state estimation unit 120 has described the example of having the second estimation unit 122, but the state estimation unit 120 may not have the second estimation unit 122. In this case, the output unit 130 may not output the information according to the estimation result of the second estimation unit 122, or may output only the information according to the estimation result of the first estimation unit 121.

For example, in the above embodiment, the output unit 130 has described an example of outputting information to the maintenance person, but it is not necessary to output the information to the maintenance person. For example, the output unit 130 may output information only to the equipment connected to the state estimation device 100 or the device recognizable by the maintenance person in the state estimation device 100.

For example, the inspection of the head 310 may be performed on the component mounting device 300, or the head 310 may be removed from the component mounting device 300 to supply power and air to the device for inspection of the head 310 (for example, the head 310). It may be done on a device that can.

For example, the present disclosure can be realized not only as a state estimation device 100, but also as a state estimation method including steps (processes) performed by each component constituting the state estimation device 100.

FIG. 10 is a flowchart showing an example of a state estimation method according to another embodiment.

The state estimation method includes a head having a head main body to which a nozzle configured to hold the object is mounted, and manages the state of the head in equipment that holds the object by the head and performs a predetermined work. As shown in FIG. 10, it is a state estimation method, and as shown in FIG. 10, an equipment log of a head having a nozzle mounted on a head body is acquired, and is based on a data set consisting of a plurality of characteristic values of the head included in the equipment log. The threshold value generation step (step S11) for generating the first threshold value and the characteristic value of the head in which the nozzle is attached to the head body and are in operation are acquired, and based on the characteristic value of the head and the first threshold value. The dataset comprises a plurality of characteristic values of the head that exceed the reference value, including a state estimation step (step S12) for estimating whether or not the head is in a state related to the malfunction.

For example, the steps in the state estimation method may be performed by a computer (computer system). Then, the present disclosure can be realized as a program for causing a computer to execute the steps included in the state estimation method. Further, the present disclosure can be realized as a non-temporary computer-readable recording medium such as a CD-ROM in which the program is recorded.

For example, when the present disclosure is realized by a program (software), each step is executed by executing the program using hardware resources such as a computer CPU, memory, and input / output circuit. .. That is, each step is executed by the CPU acquiring data from the memory or the input / output circuit or the like and performing an operation, or outputting the operation result to the memory or the input / output circuit or the like.

Further, each component included in the state estimation device 100 of the above embodiment may be realized as a dedicated or general-purpose circuit.

Further, each component included in the state estimation device 100 of the above embodiment may be realized as an LSI (Large Scale Integration) which is an integrated circuit (IC: Integrated Circuit).

Further, the integrated circuit is not limited to the LSI, and may be realized by a dedicated circuit or a general-purpose processor. A programmable FPGA (Field Programmable Gate Array) or a reconfigurable processor in which the connections and settings of the circuit cells inside the LSI can be reconfigured may be utilized.

Further, if an integrated circuit technology that replaces an LSI appears due to advances in semiconductor technology or another derived technology, it is natural that the technology will be used to integrate each component included in the state estimation device 100. You may.

In addition, a form obtained by applying various modifications to the embodiment that a person skilled in the art can think of, and a form realized by arbitrarily combining the components and functions in each embodiment within the scope of the purpose of the present disclosure. Is also included in this disclosure.

The present disclosure can be used, for example, for the management of equipment or the like that holds an object by a head and performs a predetermined work.

1 Equipment maintenance system 2 Mounting system 11 Base 12 Board transfer mechanism 13 Head movement mechanism 13a Plate 16 Tape feeder 16a Tape feed mechanism 16b Parts supply port 17 Cart 18 Carrier tape 19 Reel 20 Board recognition camera 21 Parts recognition camera 100 State estimation device 110 Threshold generation unit 120 State estimation unit 121 First estimation unit 122 Second estimation unit 123 Factor estimation unit 130 Output unit 300 Parts mounting device 310 Head 311 Head body unit 312 Nozzle 313 Nozzle holder 314 Blow valve 315 Vacuum valve 316 Common air path 317 Blow air path 318 Vacuum air path 319 Air source 320 Sensor 330 Control unit 340 Vacuum sensor

Claims (15)

A state estimation in which a head having a head main body portion to which a nozzle configured to hold the object is mounted is provided, and the state of the head is managed by the head in equipment for holding the object and performing a predetermined work. It ’s a device,
A threshold generation unit that acquires the equipment log of the head to which the nozzle is mounted and generates a first threshold value based on a data set including a plurality of characteristic values of the head included in the equipment log.
Whether or not the head is in a state related to malfunction based on the acquired characteristic value of the head and the first threshold value obtained by acquiring the characteristic value of the head in which the nozzle is attached and is in an operating state. Equipped with a state estimation unit that estimates
The data set consists of a plurality of characteristic values of the head that exceed the reference value.
State estimator. The threshold generation unit allocates the amount of state change of the head calculated by comparing a plurality of characteristic values of the head included in the equipment log with the reference value to a plurality of clusters as learning data, and the plurality of units. The first threshold value is generated based on the characteristic value of the head in one cluster having the smallest amount of state change among the clusters.
The state estimation device according to claim 1. The threshold generation unit further acquires an inspection log of the head main body to which the nozzle is not mounted, and generates a second threshold value based on a plurality of characteristic values of the head main body included in the inspection log. ,
When it is estimated that the head is in a state related to the malfunction, the state estimation unit acquires the inspection result of the head main body portion to which the nozzle is not mounted, and sets the inspection result and the second threshold value. Based on this, it is estimated whether the nozzle is in a malfunctioning state or the head main body is in a malfunctioning state.
The state estimation device according to claim 1 or 2. The plurality of characteristic values of the head main body portion are data measured by removing the nozzle from the head whose characteristic value of the head to which the nozzle is mounted exceeds the reference value.
The threshold generation unit is subjected to linear regression using learning data of a plurality of characteristic values of the head to which the nozzle is mounted and a plurality of characteristic values of the head main body corresponding to the plurality of characteristic values. Generate the second threshold based on one threshold,
The state estimation device according to claim 3. When the inspection result of the head main body exceeds the second threshold value, the state estimation unit estimates that the head main body is in a malfunctioning state.
The state estimation device according to claim 4. The state estimation unit estimates that the nozzle is in a malfunctioning state when the inspection result of the head main body portion does not exceed the second threshold value.
The state estimation device according to claim 4 or 5. The state estimation unit is generated by correlation analysis using learning data of a factor estimation value calculated from the inspection log and a factor of the head body malfunction when it is estimated that the head body is in a malfunctioning state. Using the malfunction factor estimation table, the malfunction factor of the head body is estimated based on the factor estimation value calculated from the inspection result of the head body.
The state estimation device according to any one of claims 3 to 6. The factor estimation value calculated from the inspection result of the head main body is calculated based on the statistical value of the inspection log determined that the head main body is in a normal state.
The state estimation device according to claim 7. The characteristic value is a measurement result of a flow rate or a pressure in an air path of the head main body when air is supplied from the air source of the head main body at a positive pressure or a negative pressure.
The state estimation device according to any one of claims 1 to 8. An output unit that outputs information based on the estimation result of the state estimation unit is further provided.
The state estimation device according to any one of claims 1 to 9. The information includes information indicating that the head is in a malfunction-related state or that the head is in a malfunction state.
The state estimation device according to claim 10. The information includes maintenance content for the head, which is presumed to be in a bad condition.
The state estimation device according to claim 10 or 11. The information is output to the equipment maintenance planning unit that manages the maintenance plan for maintaining the equipment.
The state estimation device according to any one of claims 10 to 12. A head having a head body to which a nozzle configured to hold an object is mounted,
A measuring unit that measures the characteristic value of the head,
The state estimation device according to any one of claims 1 to 13 for managing the state of the head.
Implementation system. A state estimation in which a head having a head main body portion to which a nozzle configured to hold the object is mounted is provided, and the state of the head is managed by the head in equipment for holding the object and performing a predetermined work. It ’s a method,
A threshold generation step of acquiring an equipment log of the head in which the nozzle is mounted on the head main body and generating a first threshold value based on a data set including a plurality of characteristic values of the head included in the equipment log. When,
The nozzle is attached to the head body to acquire the characteristic value of the head in an operating state, and the head is in a state related to malfunction based on the characteristic value of the head and the first threshold value. Including a state estimation step to estimate whether or not
The data set consists of a plurality of characteristic values of the head that exceed the reference value.
State estimation method.

Images