JP7474580B2 - Prediction system and method - Google Patents

Description

This disclosure relates to a prediction system and method for predicting the ejection results of a liquid ejection head in a liquid ejection recording device.

In a liquid jet recording device, the liquid jet head is equipped with a laminate of multiple plates, including an ejection hole plate in which many ejection holes are formed, and is configured to eject liquid droplets from each ejection hole onto a recording medium. In such a liquid jet recording device, an ejection inspection is performed in which droplets are ejected from the liquid jet head, and the ejection angle and droplet volume of the droplets are measured, and the liquid jet head is inspected based on the measurement results. A method for measuring the droplet volume is disclosed, for example, in Patent Document 1.

JP 2016-191636 A

However, in the discharge inspection, liquid ejection heads whose measurement results do not meet certain standards are discarded as defective products. Therefore, the disposal results in a loss of assembly costs. Furthermore, the discharge inspection requires time and labor costs. Therefore, it is desirable to provide a prediction system and prediction method that can predict the discharge results of a liquid ejection head at a stage prior to the discharge inspection.

A prediction system according to an embodiment of the present disclosure is a system for predicting the ejection result of each prediction jet hole of a prediction liquid jet head that ejects droplets from a plurality of prediction jet holes. This prediction system includes an appearance image acquisition unit and a prediction data acquisition unit. The appearance image acquisition unit acquires a prediction appearance image including a peripheral portion of the prediction jet hole corresponding to each prediction jet hole. The prediction data acquisition unit inputs the prediction appearance image to a learning model, thereby acquiring prediction data of the ejection result of the prediction liquid jet head from the learning model. The appearance image acquisition unit acquires, as the prediction appearance image, an image of a region of an actuator plate that controls the supply of liquid to the prediction jet hole, including at least a portion corresponding to the peripheral portion among the prediction jet hole and its peripheral portion.

A prediction method according to an embodiment of the present disclosure is a method for predicting the ejection result of each of the prediction nozzles of a prediction liquid ejection head that ejects droplets from a plurality of prediction nozzles. This prediction method includes the following two steps.
Acquiring a prediction appearance image including a peripheral portion of each of the prediction ejection holes, which corresponds to each of the prediction ejection holes; and acquiring prediction data of the ejection result of the prediction liquid ejection head from the learning model by inputting the prediction appearance image into the learning model.
The prediction method further includes one of the following:
As the prediction appearance image, an image of a region including at least a portion corresponding to the peripheral portion of the prediction injection hole and its peripheral portion in the actuator plate that controls the supply of liquid to the prediction injection hole is acquired.

In the prediction system and prediction method according to one embodiment of the present disclosure, the learning model is a model obtained by performing machine learning using learning appearance images including the surrounding areas of the learning nozzles corresponding to each of the learning nozzles of a learning liquid ejection head that ejects droplets from multiple learning nozzles as explanatory variables, and the ejection results of the learning liquid ejection head as objective variables.

The prediction system and prediction method according to one embodiment of the present disclosure make it possible to predict the ejection results of a liquid ejection head before an ejection inspection.

1 is an exploded perspective view illustrating an example of the configuration of a liquid jet head in a liquid jet recording apparatus. 2 is a diagram illustrating an example of a planar configuration of the liquid ejection head of FIG. 1 when viewed from the ejection hole plate side. FIG. 3 is a diagram illustrating a configuration example of a portion of a cross section taken along line VV illustrated in FIG. 2. 1 is a diagram illustrating a schematic configuration example of a prediction system according to an embodiment of the present disclosure. FIG. 5 is a diagram illustrating an example of a functional block of the prediction system of FIG. 4. FIG. 6 is a diagram illustrating an example of the image data of FIG. 5 . 6A and 6B are diagrams illustrating an example of the image data and the ejection result of FIG. 5 . 5 is a diagram illustrating an example of a procedure for predicting and evaluating the ejection performance of a liquid jet head in the prediction system of FIG. 4. FIG. 5 is a diagram illustrating a modified example of the schematic configuration of the prediction system in FIG. 4. 10 is a diagram illustrating an example of a procedure for predicting and evaluating the ejection performance of a liquid jet head in the prediction system of FIG. 9 . FIG. 5 is a diagram illustrating a modified example of the schematic configuration of the prediction system in FIG. 4.

Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. The description will be made in the following order.

1. Prediction Object 2. Embodiment 3. Modification

＜1. Prediction target＞
First, a prediction target in the prediction system according to an embodiment of the present disclosure will be described. The prediction target is, for example, a liquid jet head 100 used in a liquid jet recording apparatus, as shown in FIG. 1. The liquid jet recording apparatus is provided with one or more liquid jet heads 100. The prediction target in the prediction system according to an embodiment of the present disclosure is the ejection performance of the liquid jet head 100, and is, for example, at least one of the ejection angle, the ejection speed, and the amount of droplets ejected from the liquid jet head 100.

The liquid jet recording device is an inkjet printer that uses ink 150, which will be described later, to record (print) images, characters, etc., on recording paper as a recording medium. This liquid jet recording device is an ink circulation type inkjet printer that uses ink 150 by circulating it through a specified flow path. The liquid jet head 100 is a head that ejects (spits) droplets of ink 150 from multiple ejection holes (ejection holes H1, H2), which will be described later, onto recording paper, to record images, characters, etc.

Next, an example of the configuration of the liquid jet head 100, which is the prediction target, will be described with reference to Figures 1, 2, and 3. Figure 1 shows an example of the configuration of the liquid jet head 100 in an exploded perspective view. Figure 2 shows an example of the planar configuration of the liquid jet head 100 when viewed from the injection hole plate side. Figure 3 shows a part of a cross section (Z-X cross section) of the liquid jet head 100 at a location corresponding to line V-V shown in Figure 2.

The liquid jet head 100 ejects ink 150 from the center of a plurality of channels (channels C1 and C2) in the extension direction (Y-axis direction) described below. The liquid jet head 100 mainly comprises an ejection hole plate 110, an actuator plate 120, and a cover plate 130. The liquid jet head 100 comprises a laminate in which the ejection hole plate 110, the actuator plate 120, and the cover plate 130 are bonded together, for example, using an adhesive. The ejection hole plate 110, the actuator plate 120, and the cover plate 130 are laminated in this order along the Z-axis direction. In the following description, the cover plate 130 side along the Z-axis direction will be referred to as the upper side, and the ejection hole plate 110 side will be referred to as the lower side.

(Injection hole plate 110)
The injection hole plate 110 includes, for example, a metal substrate or a resin substrate, and is bonded to the lower surface of the actuator plate 120 by an adhesive. Examples of materials for the metal substrate used as the injection hole plate 110 include stainless steels such as SUS316L and SUS304. Examples of materials for the resin substrate used as the injection hole plate 110 include polyimide.

As shown in Figs. 1 and 2, the injection hole plate 110 has two injection hole rows (injection hole rows 111, 112) that each extend along the X-axis direction. These injection hole rows 111, 112 are arranged at a predetermined interval along the Y-axis direction. In this way, the liquid injection head 100 is a two-row type liquid injection head. The liquid injection head 100 may be a single-row type liquid injection head (having one injection hole row) or a multiple-row type liquid injection head (having three or more injection hole rows).

The injection hole row 111 has a plurality of injection holes H1 formed in a straight line at a predetermined interval along the X-axis direction. Each of these injection holes H1 penetrates the injection hole plate 110 along its thickness direction (Z-axis direction) and communicates with an ejection channel C1e in the actuator plate 120 described later, as shown in FIG. 3, for example. Specifically, as shown in FIG. 2, each injection hole H1 is formed in a row and is formed so as to be located at the center of the ejection channel C1e along the Y-axis direction. In addition, the formation pitch of the injection holes H1 along the X-axis direction is the same (same pitch) as the formation pitch of the ejection channel C1e along the X-axis direction. From the injection holes H1 in such an injection hole row 111, ink 150 supplied from the ejection channel C1e is ejected (sprayed), as will be described in detail later.

Similarly, the injection hole row 112 has a plurality of injection holes H2 arranged in a straight line at a predetermined interval along the X-axis direction. Each of these injection holes H2 also penetrates the injection hole plate 110 along its thickness direction and communicates with an ejection channel C2e in the actuator plate 120 described later. Specifically, as shown in FIG. 2, each injection hole H2 is formed in a row and is formed so as to be located at the center of the ejection channel C2e along the Y-axis direction. In addition, the formation pitch of the injection holes H2 along the X-axis direction is the same as the formation pitch of the ejection channel C2e along the X-axis direction. The ink 150 supplied from the ejection channel C2e is also ejected from the injection holes H2 in the injection hole row 112, as will be described in detail later.

The injection hole plate 110 has a metal substrate or a resin substrate on which a plurality of injection holes H1 and a plurality of injection holes H2 are provided. The metal substrate or the resin substrate has an injection side main surface 110B on which the injection holes Ha of the injection holes H1 and H2 are provided, and an inlet side main surface 110A on which the inlet holes Hb of the injection holes H1 and H2, which are larger than the injection holes Ha, are provided. Each of these injection holes H1 and H2 is a tapered through hole whose diameter gradually decreases as it extends downward.

Here, in the metal substrate or resin substrate, the periphery (spout edge region Ea) of the inlet Hb of each of the injection holes H1 and H2 includes at least the region of the metal substrate or resin substrate that faces the inlet Hb in the thickness direction of the metal substrate or resin substrate. The outlet edge region Ea is, for example, annular. The shape of the outlet edge region Ea is not limited to annular. The outlet edge region Ea may be, for example, an elliptical ring shape or a square ring shape. When the outlet edge region Ea is annular, the outer diameter of the outlet edge region Ea is smaller than the arrangement pitch of the injection holes H1 and H2. Therefore, each outlet edge region Ea is provided separately from each other on the ejection side main surface 110B.

The nozzle edge region Ea is a polished surface formed, for example, by mechanical polishing. The nozzle edge region Ea is an area polished, for example, by tape polishing. The area around the nozzle edge region Ea is either not polished at all or is polished more roughly than the nozzle edge region Ea.

(Actuator Plate 120)
The actuator plate 120 is a plate that controls the supply of ink 150 to each of the ejection holes H1, H2 in the ejection hole plate 110. The actuator plate 120 is a plate made of a piezoelectric material such as PZT (lead zirconate titanate). The actuator plate 120 is a so-called chevron type actuator formed by stacking two piezoelectric substrates whose polarization directions are different in the Z direction. The actuator plate 120 may be a so-called cantilever type actuator formed from one piezoelectric substrate whose polarization direction is set in one direction along the thickness direction (Z-axis direction). As shown in FIG. 1 and FIG. 2, the actuator plate 120 is provided with two channel rows (channel rows 121, 122) that extend along the X-axis direction. These channel rows 121, 122 are arranged at a predetermined interval along the Y-axis direction.

As shown in Figures 1 and 2, the above-mentioned channel row 121 has multiple channels C1 extending along the Y-axis direction. These channels C1 are arranged parallel to each other at a predetermined interval along the X-axis direction. Each channel C1 is defined by a driving wall Wd made of a piezoelectric body (actuator plate 120), and is a concave groove portion in cross-sectional view (see Figure 1).

Similarly, the channel row 122 has a number of channels C2 extending along the Y-axis direction. These channels C2 are arranged parallel to each other at a predetermined interval along the X-axis direction. Each channel C2 is also defined by the driving wall Wd described above, and is a concave groove portion in cross-sectional view.

As shown in Figures 1 and 2, the channel C1 includes an ejection channel C1e for ejecting ink 150 and a dummy channel C1d for not ejecting ink 150. In the channel row 121, these ejection channels C1e and dummy channels C1d are arranged alternately along the X-axis direction. Each ejection channel C1e is connected to an ejection hole H1 in the ejection hole plate 110, while each dummy channel C1d is not connected to the ejection hole H1 and is covered from below by the upper surface of the ejection hole plate 110.

Similarly, channel C2 includes ejection channel C2e for ejecting ink 150 and dummy channel C2d for not ejecting ink 150. In channel row 122, ejection channel C2e and dummy channel C2d are arranged alternately along the X-axis direction. Each ejection channel C2e is connected to ejection hole H2 in ejection hole plate 110, while each dummy channel C2d is not connected to ejection hole H2 and is covered from below by the upper surface of ejection hole plate 110.

2, the ejection channel C1e and the dummy channel C1d in the channel C1 are arranged in a staggered manner with respect to the ejection channel C2e and the dummy channel C2d in the channel C2. Therefore, in the liquid jet head 100, the ejection channel C1e in the channel C1 and the ejection channel C2e in the channel C2 are arranged in a staggered manner. As shown in FIG. 1, in the actuator plate 120, in the portion corresponding to the dummy channels C1d and C2d, a shallow groove portion Dd is formed which communicates with the outer end portions of the dummy channels C1d and C2d along the Y-axis direction.

1 and 3, the driving electrodes Ed extending along the Y-axis direction are provided on the opposing inner surfaces of the driving wall Wd. The driving electrodes Ed include a common electrode Edc provided on the inner surface facing the ejection channels C1e and C2e, and an active electrode Eda provided on the inner surface facing the dummy channels C1d and C2d. As shown in FIG. 3, the driving electrodes Ed (common electrode Edc and active electrode Eda) are formed on the inner surface of the driving wall Wd to the same depth as the driving wall Wd (the same depth in the Z-axis direction). In addition, an insulating film 120A is formed on the surface of the actuator plate 120 facing the injection hole plate 110 to prevent the driving electrode Ed and the injection hole plate 110 from being electrically short-circuited with each other. In addition, when the actuator plate 120 is of the above-mentioned cantilever type, the driving electrode Ed (common electrode Edc and active electrode Eda) is formed only up to the middle position in the depth direction (Z-axis direction) on the inner surface of the driving wall Wd.

A pair of opposing common electrodes Edc in the same ejection channel C1e (or ejection channel C2e) are electrically connected to each other at a common terminal (not shown). In addition, a pair of opposing active electrodes Eda in the same dummy channel C1d (or dummy channel C2d) are electrically isolated from each other. On the other hand, a pair of opposing active electrodes Eda across the ejection channel C1e (or ejection channel C2e) are electrically connected to each other at an active terminal (not shown).

As shown in FIG. 1, the liquid jet head 100 is equipped with a flexible printed circuit board 140 that electrically connects the drive electrodes Ed and a control unit that controls the liquid jet head 100. A wiring pattern (not shown) formed on the flexible printed circuit board 140 is electrically connected to the common terminal and active terminal described above. This allows a drive voltage to be applied to each drive electrode Ed from the control unit that controls the liquid jet head 100 via the flexible printed circuit board 140.

(Cover plate 130)
1, the cover plate 130 is disposed so as to close the channels C1, C2 (the channel rows 121, 122) in the actuator plate 120. Specifically, the cover plate 130 is bonded to the upper surface of the actuator plate 120 and has a plate-like structure.

As shown in FIG. 1, the cover plate 130 is formed with a pair of inlet side common ink chambers 131a, 132a and a pair of outlet side common ink chambers 131b, 132b. Specifically, the inlet side common ink chamber 131a and the outlet side common ink chamber 131b are each formed in an area corresponding to the channel row 121 (multiple channels C1) in the actuator plate 120. The inlet side common ink chamber 132a and the outlet side common ink chamber 132b are each formed in an area corresponding to the channel row 122 (multiple channels C2) in the actuator plate 120.

The inlet-side common ink chamber 131a is formed near the inner end along the Y-axis direction of each channel C1, and is a concave groove. In the inlet-side common ink chamber 131a, a supply slit Sa is formed in the area corresponding to each ejection channel C1e, penetrating the cover plate 130 along its thickness direction (Z-axis direction). Similarly, the inlet-side common ink chamber 132a is formed near the inner end along the Y-axis direction of each channel C2, and is a concave groove. In the inlet-side common ink chamber 132a, the above-mentioned supply slit Sa is also formed in the area corresponding to each ejection channel C2e.

As shown in FIG. 1, the outlet side common ink chamber 131b is formed near the outer end along the Y axis direction of each channel C1, and is a concave groove portion. In this outlet side common ink chamber 131b, in the area corresponding to each ejection channel C1e, a discharge slit Sb is formed that penetrates the cover plate 130 along its thickness direction. Similarly, the outlet side common ink chamber 132b is formed near the outer end along the Y axis direction of each channel C2, and is a concave groove portion. In this outlet side common ink chamber 132b, the above-mentioned discharge slit Sb is also formed in the area corresponding to each ejection channel C2e.

In this way, the inlet side common ink chamber 131a and the outlet side common ink chamber 131b are connected to each ejection channel C1e via the supply slit Sa and the exhaust slit Sb, respectively, but are not connected to each dummy channel C1d. In other words, each dummy channel C1d is blocked by the bottom of the inlet side common ink chamber 131a and the outlet side common ink chamber 131b.

Similarly, the inlet side common ink chamber 132a and the outlet side common ink chamber 132b are connected to each ejection channel C2e via the supply slit Sa and the exhaust slit Sb, respectively, but are not connected to each dummy channel C2d. In other words, each dummy channel C2d is blocked by the bottom of the inlet side common ink chamber 132a and the outlet side common ink chamber 132b.

2. Preferred embodiment
[composition]
Fig. 4 shows an example of a schematic configuration of a prediction system 1 according to an embodiment of the present invention. Fig. 5 shows an example of functional blocks of the prediction system 1 of Fig. 4. The prediction system 1 is a system that predicts the ejection results of each of the ejection holes H1, H2 of the liquid ejection head 100. In the prediction system 1, the prediction target is the liquid ejection head 100, and the prediction target is the ejection performance of the liquid ejection head 100, for example, at least one of the ejection angle, ejection speed, and droplet amount of droplets ejected from the liquid ejection head 100.

The prediction system 1 includes, for example, an information processing device 10, an imaging system 20, and a server device 30. The information processing device 10, the imaging system 20, and the server device 30 are connected via a network 40. The information processing device 10 is configured to be able to communicate with the information processing device 10 and the server device 30 via the network 40. The imaging system 20 is configured to be able to communicate with the information processing device 10 and the server device 30 via the network 40. The server device 30 is configured to be able to communicate with the information processing device 10 and the imaging system 20 via the network 40.

The network 40 is, for example, a network that communicates using a communication protocol (TCP/IP) that is commonly used on the Internet. The network 40 may be, for example, a secure network that communicates using a communication protocol unique to that network. The network 40 is, for example, the Internet, an intranet, or a local area network. The connection between the network 40 and the information processing device 10, the imaging system 20, or the server device 30 may be, for example, a wired LAN (Local Area Network) such as Ethernet (registered trademark), a wireless LAN such as Wi-Fi, or a mobile phone line.

The information processing device 10 includes, for example, a control unit 11, a memory 12, a display unit 13, an input unit 14, and a network IF (Interface) 15. The control unit 11 corresponds to a specific example of the "appearance image acquisition unit" and "prediction data acquisition unit" of the present disclosure. The control unit 11 includes a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit), and executes, for example, a program stored in the memory 12. The control unit 11 executes, for example, a program 12A stored in the memory 12. The program 12A predicts the ejection performance of the liquid ejection head 100 (for example, at least one of the ejection angle, ejection speed, and droplet volume of droplets ejected from the liquid ejection head 100), and causes the control unit 11 to execute a series of procedures for evaluating the liquid ejection head 100 based on the prediction result.

The control unit 11, for example, transmits an imaging instruction to the imaging system 20 via the network IF 15. The control unit 11, for example, receives image data (for example, image data 32A or image data Ia) from the imaging system 20 via the network IF 15 as a response to the imaging instruction. The control unit 11, for example, acquires image data including the peripheral portions of the injection holes H1 and H2 corresponding to each of the injection holes H1 and H2 as a response to the imaging instruction.

The control unit 11, for example, transmits the acquired image data 32A to the server device 30 via the network IF 15. The control unit 11, for example, transmits a predicted data generation instruction including the acquired image data Ia to the server device 30 via the network IF 15. The control unit 11, for example, receives predicted data Da of the ejection result of the liquid ejection head 100 from the server device 30 via the network IF 15 as a response to the predicted data generation instruction. In other words, the control unit 11, for example, inputs the image data Ia to the learning model 33 in the server device 30, thereby acquiring the predicted data Da of the ejection result of the liquid ejection head 100 from the learning model 33.

The control unit 11 evaluates the ejection performance of the liquid ejection head 100 (e.g., at least one of the ejection angle, ejection speed, and droplet volume of the droplets ejected from the liquid ejection head 100) based on, for example, the received prediction data Da. The control unit 11, for example, causes the display unit 13 to display the evaluation result.

The memory 12 stores programs executed by the control unit 11. The memory 12 is composed of, for example, a RAM (Random Access Memory), a ROM (Read Only Memory), an auxiliary storage device (such as a hard disk), etc. The memory 12 stores the program 12A.

The display unit 13 is composed of a display device such as a liquid crystal panel or an organic EL (Electro Luminescence) panel. The display unit 13 displays an image based on a video signal from the control unit 11 on the display surface 10A. The input unit 14 receives instructions from the outside (e.g., a user) and outputs the received instructions to the control unit 11. The input unit 14 is composed of, for example, a keyboard and a mouse. The input unit 14 may be composed of a touch panel provided on the display surface 10A of the information processing device 10. The network IF 15 is a communication interface for communicating with the imaging system 20 or the server device 30 via the network 40.

The server device 30 includes, for example, a control unit 31, a memory 32, a learning model 33, and a network IF 34. The control unit 31 includes a CPU and a GPU, and executes, for example, a program stored in the memory 32. The control unit 31 receives, for example, an instruction to generate predicted data Da including image data Ia obtained by the imaging system 20 from the imaging system 20 via the network IF 34, and inputs the received image data Ia to the learning model 33. The control unit 31 acquires predicted data Da of the ejection result from the learning model 33 in response to the input of the image data Ia, and transmits the acquired predicted data Da to the information processing device 10 via the network IF 34.

The network IF 34 is a communication interface for communicating with the information processing device 10 or the imaging system 20 via the network 40. The memory 32 stores programs executed by the control unit 31, etc. The memory 32 is composed of, for example, a RAM, a ROM, an auxiliary storage device (such as a hard disk), etc. The memory 32 further stores a plurality of image data 32A and a plurality of ejection results 32B. The plurality of image data 32A and the plurality of ejection results 32B are data obtained from an ejection hole plate before being incorporated into a learning liquid ejection head described below, an actuator plate before being incorporated into a learning liquid ejection head, or a learning liquid ejection head. The plurality of image data 32A and the plurality of ejection results 32B are used when learning the learning model 33.

The image data 32A is composed of gradation data for each pixel of the image, for example, as shown in FIG. 6. The image data 32A is composed of gradation data for each pixel of the image, for example, as shown in FIG. 7. FIG. 7 conceptually shows how the image data 32A and the ejection result 32B are stored in the memory 32 in correspondence with each imaging target (injection hole). The image data 32A is not binarized data, but grayscale data. Therefore, in the image data 32A, the portion corresponding to the injection hole is low-gradation data equivalent to approximately black, and the peripheral portion of the injection hole (for example, the nozzle edge region Ea) is higher-gradation data than the portion corresponding to the injection hole. The image data 32A may be colored data. Note that the image data Ia is also data similar to the image data 32A, and is not binarized data, but grayscale data or colored data. Therefore, in the image data Ia, the portion corresponding to the injection hole is low-gradation data equivalent to approximately black, and the peripheral portion of the injection hole (for example, the nozzle edge region Ea) is higher-gradation data than the portion corresponding to the injection hole. In FIG. 7, the ejection result 32B is the ejection angle of the droplet. Note that the ejection result 32B may also be the ejection speed of the droplet or the amount of the droplet.

The learning model 33 is a model obtained by performing machine learning using image data 32A (learning appearance image) as an explanatory variable and ejection result 32B (ejection result of the learning liquid ejection head) as a target variable. The learning liquid ejection head has a common configuration with the above-mentioned liquid ejection head 100. Like the above-mentioned liquid ejection head 100, the learning liquid ejection head mainly comprises an ejection hole plate 110, an actuator plate 120, and a cover plate 130, and these ejection hole plate 110, actuator plate 120, and cover plate 130 comprise a laminate that is bonded to each other using, for example, an adhesive or the like.

Image data 32A (learning appearance image) is data obtained by imaging an imaging area α of a nozzle plate before being incorporated into a learning liquid ejection head, an actuator plate before being incorporated into a learning liquid ejection head, a nozzle plate with a learning ejection hole formed therein and an actuator plate that controls the supply of liquid to the learning ejection hole, which are bonded together, or a learning liquid ejection head, using imaging system 20. Discharge result 32B (discharge result of learning liquid ejection head) is actual measurement data obtained by performing a discharge inspection on an assembled learning liquid ejection head. Discharge inspection refers to measuring at least one of the discharge angle, discharge speed, and droplet volume of droplets discharged from an assembled learning liquid ejection head.

When image data Ia (prediction appearance image) is input, the learning model 33 generates (predicts) prediction data Da of the ejection result of the liquid ejection head 100 (prediction liquid ejection head) based on the learning results, and outputs the generated (predicted) prediction data Da.

The image data Ia (prediction appearance image) is data obtained by imaging the injection hole plate 110 before being incorporated into the liquid injection head 100, the actuator plate 120 before being incorporated into the liquid injection head 100, the injection hole plate 110 and the actuator plate 120 which are bonded together, or the imaging area α of the liquid injection head 100 using the imaging system 20. The prediction data Da (prediction data of the ejection result of the prediction liquid injection head) is prediction data for the ejection result that would be obtained if an ejection inspection was performed on the assembled liquid injection head 100. The ejection inspection refers to measuring at least one of the ejection angle, ejection speed, and droplet volume of the droplets ejected from the assembled liquid injection head 100. The prediction data Da is data obtained by the learning model 33 before the liquid injection head 100 itself is assembled, or before the liquid injection head 100 is incorporated into the liquid injection recording device.

The imaging system 20 includes, for example, a camera 21, a light source unit 22, a half mirror 23, an XY stage 24, and a control unit 25.

The camera 21 is configured to include, for example, a CCD (Charge Coupled Device) image sensor or a CMOS (Complementary metal-oxide-semiconductor) image sensor. The camera 21 captures an image, for example, under the control of the control unit 25, and outputs the image (image data) obtained thereby to the control unit 25. The camera 21 captures an image of the imaging area α, and outputs the image data obtained thereby (for example, image data 32A or image data Ia) to the control unit 25.

The imaging area α is, for example, as shown in FIG. 2, an area including at least the peripheral portion (ejection outlet edge area Ea) of the ejection outlet Ha of the ejection hole H1 or the ejection hole H2 and its peripheral portion (ejection outlet edge area Ea) on the ejection side main surface 110B of the ejection hole plate 110. The imaging area α is, for example, as shown in FIG. 2, an area including the ejection outlet Ha of the ejection hole H1 or the ejection hole H2 and the peripheral portion (ejection outlet edge area Ea) of the ejection hole plate 110. At this time, the image data (image data 32A or image data Ia) is image data of an area including the ejection outlet Ha of the ejection hole H1 or the ejection hole H2 and the peripheral portion (ejection outlet edge area Ea) of the ejection hole plate 110 on the ejection side main surface 110B of the ejection hole plate 110. However, the imaging area α does not include the nozzle Ha and nozzle edge area Ea of the adjacent injection hole H1 or injection hole H2, for example, as shown in FIG. 2. When imaging the imaging area α of the injection hole plate 110 in the imaging system 20, the injection hole plate 110 is imaged before being incorporated into the liquid injection head 100.

The imaging area α may be, for example, an area including at least the peripheral portion of the nozzle Ha of the injection hole H1 or the injection hole H2 when assembled on the surface of the actuator plate 120 facing the injection hole plate 110 and the peripheral portion thereof. The imaging area α is, for example, an area including an area facing the nozzle edge area Ea of the nozzle Ha of the injection hole H1 or the injection hole H2 when assembled on the surface of the actuator plate 120 facing the injection hole plate 110, and an area including an area facing the nozzle Ha of the injection hole H1 or the injection hole H2 when assembled on the surface of the actuator plate 120 facing the injection hole plate 110. At this time, the image data (image data 32A or image data Ia) is image data of an area including an area facing the nozzle edge area Ea of the nozzle Ha of the injection hole H1 or the injection hole H2 when assembled on the surface of the actuator plate 120 facing the injection hole plate 110, and an area including an area facing the nozzle Ha of the injection hole H1 or the injection hole H2 when assembled. However, the imaging region α of the actuator plate 120 includes a portion of the ejection channel C1e (groove portion) or the ejection channel C2e (groove portion), but does not include the dummy channel C1d (groove portion) or the dummy channel C2d (groove portion). When the imaging system 20 images the imaging region α of the actuator plate 120, the actuator plate 120 is imaged before being incorporated into the liquid ejection head 100.

When the injection hole plate 110 is configured to include a light-transmitting resin substrate, the imaging area α may be, for example, an area facing the peripheral portion of the nozzle Ha of the injection hole H1 or the injection hole H2 in the injection hole plate 110 and the actuator plate 120, which are bonded together, and an area facing the nozzle Ha of the injection hole H1 or the injection hole H2. In this case, the image data (image data 32A or image data Ia) is image data of an area including an area facing the peripheral portion of the nozzle Ha of the injection hole H1 or the injection hole H2 in the injection hole plate 110 and the actuator plate 120, which are bonded together, and an area facing the nozzle Ha of the injection hole H1 or the injection hole H2. However, the imaging area α in the injection hole plate 110 and the actuator plate 120, which are bonded together, includes an area facing a part of the ejection channel C1e (groove portion) or the ejection channel C2e (groove portion). The imaging area α of the ejection hole plate 110 and the actuator plate 120, which are bonded together, may or may not include an area facing the dummy channel C1d (groove portion) or the dummy channel C2d (groove portion). When imaging the imaging area α of the ejection hole plate 110 and the actuator plate 120, which are bonded together, in the imaging system 20, the ejection hole plate 110 and the actuator plate 120, which are bonded together, are imaged before being incorporated into the liquid ejection recording device.

When the injection hole plate 110 is configured to include a light-transmitting resin substrate, the imaging area α may be, for example, an area facing the peripheral portion of the nozzle Ha of the injection hole H1 or the injection hole H2 in the injection hole plate 110 and the actuator plate 120 incorporated in the liquid injection head 100, and an area facing the nozzle Ha of the injection hole H1 or the injection hole H2, as shown in FIG. 3. In this case, the image data (image data 32A or image data Ia) is image data of an area including an area facing the peripheral portion of the nozzle Ha of the injection hole H1 or the injection hole H2 in the injection hole plate 110 and the actuator plate 120 incorporated in the liquid injection head 100, and an area facing the nozzle Ha of the injection hole H1 or the injection hole H2. However, the imaging area α of the liquid injection head 100 includes an area facing a part of the ejection channel C1e (groove portion) or the ejection channel C2e (groove portion). The imaging area α of the liquid jet head 100 may or may not include an area facing the dummy channel C1d (groove portion) or the dummy channel C2d (groove portion). When the imaging system 20 images the imaging area α of the liquid jet head 100, the liquid jet head 100 is imaged before being incorporated into the liquid jet recording device.

The light source unit 22 emits light for illuminating the imaging area α. The light source unit 22 includes, for example, a mercury lamp and an LED. The half mirror 23 is an optical component that reflects light from the light source unit 22 and transmits the reflected light in the imaging area α. The half mirror 23 is, for example, an optical component in which a dielectric multilayer film is laminated on the surface of a glass substrate. The XY stage 24 moves the injection hole plate 110, the actuator plate 120, or the liquid injection head 100 in the XY plane (within a plane having a normal parallel to the optical axis of the camera 21) according to the control of the control unit 25, thereby moving the imaging area α in the injection hole plate 110, the actuator plate 120, or the liquid injection head 100. The control unit 25 instructs the camera 21 to capture an image or instructs the XY stage 24 to move the imaging target area according to the control of the information processing device 10. The control unit 25 transmits the image data output from the camera 21 (e.g., image data 32A or image data Ia) to the information processing device 10 via the network 40.

[motion]
Next, the operation of the prediction system 1 according to an embodiment of the present disclosure will be described.

Figure 8 shows an example of a procedure for predicting and evaluating the ejection performance of the liquid ejection head 100 in the prediction system 1. First, the information processing device 10 instructs the imaging system 20 to capture an image. The imaging system 20 then controls the XY stage 24 so that the injection hole plate 110, the actuator plate 120, the injection hole plate 110 and the actuator plate 120 bonded together, or one of the many imaging target areas in the liquid ejection head 100, moves to the imaging area α on the XY stage 24. Next, the imaging system 20 captures the imaging area α and transmits the image data Ia obtained thereby to the information processing device 10 via the network 40.

The imaging system 20 controls the XY stage 24 so that the remaining imaging target areas among the multiple imaging target areas on the XY stage 24, the injection hole plate 110, the actuator plate 120, the injection hole plate 110 and the actuator plate 120 bonded together, or the liquid injection head 100, move sequentially to the imaging area α each time an image is captured. The imaging system 20 captures the imaging area α each time the imaging target area is moved, and transmits the image data Ia obtained thereby to the information processing device 10 via the network 40.

When the information processing device 10 acquires a plurality of image data Ia corresponding to all imaging target areas, it generates a predicted data generation instruction including the image data Ia for each acquired image data Ia, and transmits the generated predicted data generation instruction to the server device 30 via the network 40. When the server device 30 acquires the predicted data generation instruction, it inputs the image data Ia included in the acquired predicted data generation instruction to the learning model 33. When the image data Ia is input, the learning model 33 generates (predicts) predicted data Da of the ejection result of the injection hole plate 110, the actuator plate 120 on the XY stage 24, or the liquid injection head 100 incorporating the injection hole plate 110 and the actuator plate 120 bonded to each other, or the liquid injection head 100 on the XY stage 24, based on the learning result, and outputs the generated (predicted) predicted data Da to the control unit 31.

The server device 30 transmits the acquired predicted data Da to the information processing device 10 via the network 40. When the information processing device 10 receives the predicted data Da, it evaluates the received predicted data Da. Specifically, the information processing device 10 determines whether the received predicted data Da satisfies a certain criterion. As a result, if the predicted data Da satisfies the certain criterion, the information processing device 10 displays an image including the character string "Passed Criterion" (evaluation result) on the display unit 13. On the other hand, if the predicted data Da does not satisfy the certain criterion, the information processing device 10 displays an image including the character string "Failed Criterion" (evaluation result) on the display unit 13. In this way, the prediction system 1 predicts and evaluates the ejection performance of the liquid ejection head 100.

[effect]
Next, effects of the prediction system 1 according to an embodiment of the present disclosure will be described.

In this embodiment, image data Ia including the peripheral areas of the injection holes H1 and H2 corresponding to each of the injection holes H1 and H2 is input to the learning model 33, and prediction data Da of the ejection results of the liquid ejection head 100 is obtained from the learning model 33. This makes it possible to obtain prediction data Da based on information on the peripheral areas of the injection holes H1 and H2, which can have a significant effect on the ejection performance. As a result, the ejection performance of the liquid ejection head 100 can be accurately evaluated. This makes it possible to eliminate the loss of assembly costs associated with disposal, and the time and labor costs required for ejection inspection.

In this embodiment, an image including the peripheral areas of the ejection holes H1, H2 in the ejection hole plate 110 before it is assembled into the liquid ejection head 100 is acquired as image data Ia from the imaging system 20. This makes it possible to reduce the loss of costs for assembling the liquid ejection head 100 itself and the costs for assembling the liquid ejection head 100 into the liquid ejection recording device that would occur if an ejection inspection were performed on the liquid ejection head 100, even if the prediction system 1 fails the criteria. In addition, it is possible to eliminate the time and labor costs required for the ejection inspection.

In this embodiment, as image data Ia, an image including the peripheral areas of the ejection holes H1, H2 in the actuator plate 120 before it is assembled into the liquid ejection head 100 is acquired from the imaging system 20. This makes it possible to reduce the loss of assembly costs for the liquid ejection head 100 itself and the assembly costs for the liquid ejection head 100 into the liquid ejection recording device that would occur if an ejection inspection were performed on the liquid ejection head 100, even if the prediction system 1 fails the criteria. In addition, it is possible to eliminate the time and labor costs required for the ejection inspection.

In this embodiment, an image including peripheral portions of the ejection holes H1, H2 in the ejection hole plate 110 and the actuator plate 120, which are bonded together, is acquired as image data Ia from the imaging system 20. This makes it possible to acquire images from both the ejection hole plate 110 and the actuator plate 120, thereby increasing input information and enabling more accurate evaluation. Furthermore, even if the prediction system 1 fails the criteria, it is possible to suppress loss of costs for assembling the liquid ejection head 100 into the liquid ejection recording apparatus, which would have occurred if an ejection inspection had been performed on the liquid ejection head 100.

In this embodiment, as the image data Ia, an image including peripheral portions of the ejection holes H1, H2 in the ejection hole plate 110 and the actuator plate 120 incorporated in the liquid ejection head 100 is acquired from the imaging system 20. This makes it possible to acquire images from both the ejection hole plate 110 and the actuator plate 120, thereby increasing the amount of input information and making the evaluation more accurate. Furthermore, even if the prediction system 1 fails the criteria, it is possible to suppress loss of costs for assembling the liquid ejection head 100 into the liquid ejection recording apparatus, which would have occurred if an ejection inspection had been performed on the liquid ejection head 100.

In this embodiment, a grayscale or colored image is acquired as image data Ia by the imaging system 20. This allows the image data Ia, which includes minute differences in surface shape that appear in the shading and color of the image, to be used as input data for the learning model 33. As a result, the ejection performance of the liquid ejection head 100 can be accurately predicted and evaluated.

In this embodiment, the ejection result 32B as the objective variable is at least one of the measured data of the droplet ejection angle, ejection speed, and droplet volume, obtained by performing an ejection test on the learning liquid ejection head. In addition, the predicted data Da obtained from the learning model 33 is predicted data for at least one of the droplet ejection angle, ejection speed, and droplet volume, which would be obtained if an ejection test was performed on the liquid ejection head 100. This makes it possible to accurately evaluate the ejection performance of the liquid ejection head 100 without performing an ejection test on the liquid ejection head 100.

In this embodiment, the information processing device 10 may display the predicted data Da on the display unit 13. Also, in this embodiment, the information processing device 10 may output at least one of the predicted data Da and the evaluation result to an external device.

3. Modifications
Although the present disclosure has been described above by way of the embodiment, the present disclosure is not limited to this embodiment and various modifications are possible.

[Variation A]
Fig. 9 shows a modified example of the functional blocks of the prediction system 1 according to the embodiment. In the embodiment, for example, as shown in Fig. 9, the server device 30 may be omitted. In this case, the information processing device 10 may have, for example, a memory 32 that stores a plurality of image data 32A and a plurality of ejection results 32B, and a learning model 33, as shown in Fig. 9.

Figure 10 shows an example of a procedure for predicting and evaluating the ejection performance of the liquid ejection head 100 in the prediction system 1 according to this modified example. First, the information processing device 10 instructs the imaging system 20 to capture an image. The imaging system 20 then controls the XY stage 24 so that one of the many imaging target areas on the injection hole plate 110, the actuator plate 120, or the liquid ejection head 100 on the XY stage 24 moves to the imaging area α. Next, the imaging system 20 captures the imaging area α and transmits the image data Ia obtained thereby to the information processing device 10 via the network 40.

The imaging system 20 controls the XY stage 24 so that the remaining imaging target areas among the multiple imaging target areas on the injection hole plate 110, the actuator plate 120, or the liquid injection head 100 on the XY stage 24 are moved sequentially to the imaging area α each time an image is captured. The imaging system 20 captures the imaging area α each time the imaging target area is moved, and transmits the image data Ia obtained thereby to the information processing device 10 via the network 40.

When the information processing device 10 acquires a plurality of image data Ia corresponding to all imaging target areas, it inputs the acquired image data Ia to the learning model 33. When the image data Ia is input, the learning model 33 generates (predicts) predicted data Da of the ejection result of the liquid ejection head 100 incorporating the ejection hole plate 110, the actuator plate 120, or the ejection hole plate 110 and the actuator plate 120 bonded together on the XY stage 24, or the liquid ejection head 100 on the XY stage 24, based on the learning result, and outputs the generated (predicted) predicted data Da to the control unit 11.

The information processing device 10 evaluates the acquired predicted data Da. Specifically, the information processing device 10 determines whether the acquired predicted data Da meets a certain standard. As a result, if the predicted data Da meets the certain standard, the information processing device 10 displays an image including the character string "Passed Standard" (evaluation result) on the display unit 13. On the other hand, if the predicted data Da does not meet the certain standard, the information processing device 10 displays an image including the character string "Failed Standard" (evaluation result) on the display unit 13. In this manner, the prediction system 1 according to this modified example predicts and evaluates the ejection performance of the liquid ejection head 100.

In the prediction system 1 according to this modified example, as in the above embodiment, image data Ia including the peripheral areas of the injection holes H1 and H2 corresponding to each of the injection holes H1 and H2 is input to the learning model 33, and prediction data Da of the ejection results of the liquid ejection head 100 is obtained from the learning model 33. This makes it possible to obtain prediction data Da based on information about the peripheral areas of the injection holes H1 and H2, which can have a significant effect on the ejection performance. As a result, the ejection performance of the liquid ejection head 100 can be accurately evaluated.

In this modified example, the information processing device 10 may display the predicted data Da on the display unit 13. In addition, in this modified example, the information processing device 10 may output at least one of the predicted data Da and the evaluation result to an external device.

[Variation B]
In the above embodiments and their variations, the multiple image data 32A (learning appearance images) may include image data obtained by using the imaging system 20 to capture an imaging area α of any of the following: an injection hole plate before it is incorporated into the learning liquid jet head, an actuator plate before it is incorporated into the learning liquid jet head, an injection hole plate and an actuator plate that are bonded to each other before they are incorporated into the learning liquid jet head, and the learning liquid jet head, rotated a predetermined angle θ in the XY plane (a plane having a normal line parallel to the optical axis of the camera 21).

It is assumed that the ejection result 32B (the ejection result of the learning liquid ejection head) is the ejection angle φ of the droplet. At this time, the ejection vector v of the droplet can be expressed by the following formula (1).
|v|(cosφ, sinφ)... (1)

Here, it is assumed that the injection hole plate before being incorporated into the liquid ejection head for learning, the actuator plate before being incorporated into the liquid ejection head for learning, the injection hole plate and the actuator plate which are attached to each other before being incorporated into the liquid ejection head for learning, or the liquid ejection head for learning is rotated by an angle θ in the XY plane (in a plane having a normal line parallel to the optical axis of the camera 21). At this time, the ejection vector R(v) of the droplet can be expressed by the following formula (2).
R(v) = |v| (cos(φ+θ), sin(φ+θ)) ... (2)

In this way, by rotating the imaged object by a predetermined angle θ to generate image data 32A, it is possible to obtain image data 32A that is different from the image data 32A before the imaged object is rotated as learning data. This makes it possible to generate a large amount of learning data even when the number of injection holes for learning is small.

In addition, in the above embodiment and its modified examples, the multiple image data 32A (learning appearance images) may include images obtained by rotating an image of any of the following: the injection hole plate before it is incorporated into the learning liquid ejection head, the actuator plate before it is incorporated into the learning liquid ejection head, the injection hole plate and the actuator plate that are attached to each other before they are incorporated into the learning liquid ejection head, and the learning liquid ejection head by a predetermined angle θ.

In this way, by rotating the acquired image by a predetermined angle θ, new image data 32A is generated, and image data 32A that is different from the image data 32A before the rotation can be acquired as learning data. This makes it possible to generate a large amount of learning data even when the number of learning injection hole plates is small. As a result, the amount of learning image data 32A can be increased without increasing the imaging time.

In this modification, the control unit 11 performs a predetermined process on the image data 32A so that the image data 32A before rotation and the image data 32A after rotation are treated as different data as learning data when the learning model 33 is trained. For example, when the learning model 33 is trained, the control unit 11 converts the image data 32A into series data in which all pixel data contained in the image data 32A are arranged in a single line as shown in FIG. 7, and inputs the series data obtained thereby to the learning model 33. For example, when the learning model 33 is trained, the control unit 11 performs a predetermined process on the image data 32A for each of a plurality of adjacent pixel data, and inputs the convolution data obtained thereby to the learning model 33. In this way, in this modification, when the learning model 33 is trained, the above-mentioned process is performed on the image data 32A, so that a plurality of image data 32A that are different as learning data can be generated from one image data 32A. Therefore, even if the number of injection hole plates for training is small, a large amount of learning data can be generated.

[Variation C]
In the imaging system 20 according to the above embodiment and its modified example, the half mirror 23 may be omitted. In that case, however, the light source unit 22 may be configured as an annular light source with the optical axis of the camera 21 as the central axis. In this way, light having a more uniform light intensity distribution can be irradiated onto the imaging region α, so that the influence on contrast in the image data (image data 32A, image data Ia) caused by non-uniformity in the light intensity distribution of the light source unit 22 can be reduced.

[Modification D]
Fig. 11 shows one modified example of the prediction system 1 according to the modified example A. In the modified example A, for example, as shown in Fig. 11, a control unit 25 that controls the camera 21 and the XY stage 24 included in the imaging system 20 may be provided in the information processing device 10. In this case, the control unit 25 in the information processing device 10 controls the imaging system 20 without going through the network 40. Even in this case, the same effect as the modified example can be obtained.

[Modification E]
In the above embodiment and its modified examples, the imaging region α may be a region that includes the peripheral portion (jet port edge region Ea) of the ejection port Ha of the ejection port H1 or the ejection port H2, but does not include the ejection port Ha of the ejection port H1 or the ejection port H2, in the ejection-side main surface 110B of the ejection port plate 110. In this case, the image data (image data 32A or image data Ia) is image data of a region that includes the peripheral portion (jet port edge region Ea) of the ejection port Ha of the ejection port H1 or the ejection port H2, but does not include the ejection port Ha of the ejection port H1 or the ejection port H2, in the ejection-side main surface 110B of the ejection port plate 110.

Even in this case, it is possible to obtain predicted data Da based on information about the areas surrounding the injection holes H1 and H2, which can have a significant effect on the ejection performance. As a result, it is possible to accurately evaluate the ejection performance of the liquid ejection head 100. This makes it possible to eliminate the loss of assembly costs associated with disposal, as well as the time and labor costs required for ejection inspection.

In the above embodiment and its modified examples, the imaging area α may be an area on the surface of the actuator plate 120 facing the injection hole plate 110 that faces the peripheral portion of the nozzle Ha of the injection hole H1 or injection hole H2 when assembled, but does not include an area facing the nozzle Ha of the injection hole H1 or injection hole H2 when assembled. In this case, the image data (image data 32A or image data Ia) is image data of an area on the surface of the actuator plate 120 facing the injection hole plate 110 that includes an area facing the peripheral portion of the nozzle Ha of the injection hole H1 or injection hole H2 when assembled, but does not include an area facing the nozzle Ha of the injection hole H1 or injection hole H2 when assembled.

Even in this case, it is possible to obtain predicted data Da based on information about the area facing the peripheral parts of the injection holes H1 and H2, which can have a significant effect on the ejection performance. As a result, it is possible to accurately evaluate the ejection performance of the liquid ejection head 100. This makes it possible to eliminate the loss of assembly costs associated with disposal, as well as the time and labor costs required for ejection inspection.

In the above embodiment and its modified example, when the injection hole plate 110 is configured to include a light-transmitting resin substrate, the imaging area α may include an area facing the peripheral portion of the injection hole H1 or the injection hole H2 nozzle Ha in the injection hole plate 110 and the actuator plate 120 that are bonded together, and may not include an area facing the injection hole H1 or the injection hole H2 nozzle Ha in the injection hole plate 110 and the actuator plate 120 that are bonded together. In this case, the image data (image data 32A or image data Ia) is image data of an area including an area facing the peripheral portion of the injection hole H1 or the injection hole H2 nozzle Ha in the injection hole plate 110 and the actuator plate 120 that are bonded together, and not including an area facing the injection hole H1 or the injection hole H2 nozzle Ha in the injection hole plate 110 and the actuator plate 120 that are bonded together.

Even in this case, it is possible to obtain predicted data Da based on information about the area facing the peripheral parts of the injection holes H1 and H2, which can have a significant effect on the ejection performance. As a result, it is possible to accurately evaluate the ejection performance of the liquid ejection head 100. This makes it possible to eliminate the loss of assembly costs associated with disposal, as well as the time and labor costs required for ejection inspection.

In the above embodiment and its modified example, when the injection hole plate 110 is configured to include a light-transmitting resin substrate, the imaging area α may include an area facing the peripheral portion of the nozzle Ha of the injection hole H1 or the injection hole H2 in the injection hole plate 110 and the actuator plate 120 incorporated in the liquid injection head 100, but may not include an area facing the nozzle Ha of the injection hole H1 or the injection hole H2 in the injection hole plate 110 and the actuator plate 120 incorporated in the liquid injection head 100. In this case, the image data (image data 32A or image data Ia) is image data of an area including an area facing the peripheral portion of the nozzle Ha of the injection hole H1 or the injection hole H2 in the injection hole plate 110 and the actuator plate 120 incorporated in the liquid injection head 100, but not including an area facing the nozzle Ha of the injection hole H1 or the injection hole H2 in the injection hole plate 110 and the actuator plate 120 incorporated in the liquid injection head 100.

Even in this case, it is possible to obtain predicted data Da based on information about the area facing the peripheral parts of the injection holes H1 and H2, which can have a significant effect on the ejection performance. As a result, it is possible to accurately evaluate the ejection performance of the liquid ejection head 100. This makes it possible to eliminate the loss of assembly costs associated with disposal, as well as the time and labor costs required for ejection inspection.

[Other variations]
Furthermore, for example, in the above embodiment and its modified examples, specific configuration examples (shape, arrangement, number, etc.) of each component in the liquid jet head 100 have been given and described, but the configuration is not limited to what has been described in the above embodiment and its modified examples, and other shapes, arrangements, numbers, etc. may also be used.

For example, in the above embodiment and its modified examples, the case where the injection hole rows 111, 112 each extend linearly along the X-axis direction has been described, but this is not limited to the example, and for example, the injection hole rows 111, 112 may each extend in an oblique direction. Furthermore, the shape of the injection holes H1, H2 is not limited to the circular shape described in the above embodiment and its modified examples, and may be, for example, a polygonal shape such as a triangular shape, an elliptical shape, a star shape, or the like.

For example, in the above embodiment and its modified examples, the liquid jet head 100 is of a side shoot type, but this is not limited to this example, and for example, the liquid jet head 100 may be of another type. For example, in the above embodiment and its modified examples, the liquid jet head 100 is of a circulation type, but this is not limited to this example, and for example, the liquid jet head 100 may be of another type that does not circulate.

For example, the series of processes described in the above embodiment and its modified examples may be performed by hardware (circuits) or software (programs). When performed by software, the software is composed of a group of programs for causing a computer to execute each function. Each program may be, for example, pre-installed in the computer and used, or may be installed in the computer from a network or recording medium and used.

In addition, in the above embodiment and its modified examples, an inkjet printer has been described as one specific example of a "liquid jet recording device" in the present disclosure, but the present disclosure is not limited to this example and can be applied to devices other than inkjet printers. In other words, the "liquid jet head" and "jet hole plate" of the present disclosure may be applied to devices other than inkjet printers. Specifically, for example, the "liquid jet head" and "jet hole plate" of the present disclosure may be applied to devices such as facsimiles and on-demand printers.

In addition, in the above embodiment and its modified examples, the recording object of the liquid jet recording device is recording paper, but the recording object of the "liquid jet recording device" of the present disclosure is not limited to this. For example, letters and patterns can be formed by spraying ink onto various materials such as cardboard, cloth, plastic, and metal. Furthermore, the recording object does not have to be flat, and various three-dimensional objects such as food, building materials such as tiles, furniture, and automobiles can also be painted or decorated. Furthermore, the "liquid jet recording device" of the present disclosure can be used to print textiles, or to create three-dimensional shapes by solidifying the ink after spraying (a so-called 3D printer).

Furthermore, the various examples described above may be applied in any combination.

The effects described in this specification are merely examples and are not limiting, and other effects may also be present.

The present disclosure can also be configured as follows.
(1)
A prediction system for predicting a discharge result for each of a plurality of prediction ejection holes of a prediction liquid ejection head that ejects liquid droplets from the plurality of prediction ejection holes,
an appearance image acquisition unit that acquires a prediction appearance image including a peripheral portion of the prediction injection hole corresponding to each of the prediction injection holes;
a prediction data acquisition unit that acquires predicted data of the ejection result of the prediction liquid ejection head from the learning model by inputting the predictive appearance image into a learning model obtained by performing machine learning using a learning appearance image that corresponds to each of the learning ejection holes of a learning liquid ejection head that ejects droplets from a plurality of learning ejection holes as an explanatory variable and that includes peripheral portions of the learning ejection holes, and using the ejection result of the learning liquid ejection head as an objective variable.
(2)
The prediction system described in (1), wherein the appearance image acquisition unit acquires, as the prediction appearance image, an image including at least the surrounding portion of the prediction injection hole and its surrounding portion in an injection hole plate in which the prediction injection hole is formed before the prediction liquid injection head is assembled.
(3)
The prediction system described in (1), wherein the appearance image acquisition unit acquires, as the prediction appearance image, an image of an area in an actuator plate that controls the supply of liquid to the prediction injection hole before being incorporated into the prediction liquid injection head, the area including at least a portion corresponding to the prediction injection hole and its surrounding portion when incorporated.
(4)
The appearance image acquisition unit acquires, as the prediction appearance image, an image including at least the surrounding portion of the prediction jet hole and the surrounding portion of the jet hole plate in which the prediction jet hole is formed and the actuator plate that controls the supply of liquid to the prediction jet hole, which are pasted together.
(5)
The prediction system according to any one of claims 1 to 5, wherein the appearance image acquisition unit acquires, as the prediction appearance image, an image including a peripheral portion of the prediction jet hole in the prediction liquid jet head.
(6)
The prediction system according to any one of (1) to (5), wherein the appearance image acquisition unit acquires a grayscale or colored image as the prediction appearance image.
(7)
the ejection result as the objective variable is at least one of actual measurement data of an ejection angle, an ejection speed, and an amount of droplets, which is obtained by performing an ejection test on the learning liquid ejection head,
The prediction data obtained from the learning model is prediction data for at least one of the droplet ejection angle, ejection speed, and droplet volume that would be obtained if an ejection test was performed on the prediction liquid ejection head.
(8)
The prediction system described in any one of (1) to (7), wherein the training appearance images as explanatory variables include images of a jet hole plate in which the training jet hole is formed before being incorporated into the training liquid jet head, an actuator plate that controls the supply of liquid to the training jet hole before being incorporated into the training liquid jet head, and an image of either the jet hole plate or the actuator plate, which are bonded to each other, rotated a predetermined angle.
(9)
The prediction system described in any one of (1) to (7), wherein the training appearance images as explanatory variables include an injection hole plate in which the training injection hole is formed before being incorporated into the training liquid injection head, an actuator plate that controls the supply of liquid to the training injection hole before being incorporated into the training liquid injection head, and an image obtained by rotating, by a predetermined angle, either an image of the injection hole plate or the actuator plate that are bonded to each other.
(10)
A prediction method for predicting a discharge result for each of a plurality of prediction ejection holes of a prediction liquid ejection head that ejects liquid droplets from the plurality of prediction ejection holes, the method comprising the steps of:
acquiring a prediction appearance image including a peripheral portion of the prediction injection hole corresponding to each of the prediction injection holes;
a learning appearance image including a peripheral portion of each of the learning nozzles of a learning liquid jet head that jets droplets from a plurality of learning nozzles is used as an explanatory variable, and the prediction appearance image is input to a learning model obtained by performing machine learning using the ejection result of the learning liquid jet head as a target variable, thereby obtaining prediction data of the ejection result of the prediction liquid jet head from the learning model.

1...Prediction system 1...Information processing device, 10A...Display surface, 11...Control unit, 12...Memory, 13...Display unit, 14...Input unit, 15...Network IF, 20...Imaging system, 21...Camera, 22...Light source unit, 23...Half mirror, 24...XY stage, 25...Control unit, 30...Server device, 31...Control unit, 32...Memory, 32A, Ia...Image data, 33...Learning model, 34...Network IF, 40...Network, 100...Liquid injection head, 110A...Inlet side main surface, 110B...Ejection side main surface, 110...Injection hole plate, 120...Actuator plate , 120A...insulating film, 121, 122...channel row, 130...cover plate, 131a, 132a...inlet common ink chamber, 131b, 132b...outlet common ink chamber, 140...flexible printed circuit board, 150...ink, C1, C2...channel, C1d, C2d...dummy channel, C1e, C2e...ejection channel, Da...predicted data, Ea...nozzle edge area, Ed...driving electrode, Eda...active electrode, Edc...common electrode, Wd...driving wall, H1, H2...ejection hole, Hb...inlet, Sa...supply slit, Sb...exhaust slit, α...imaging area.

Claims (7)

A prediction system for predicting a discharge result for each of a plurality of prediction ejection holes of a prediction liquid ejection head that ejects liquid droplets from the plurality of prediction ejection holes,
an appearance image acquisition unit that acquires a prediction appearance image including a peripheral portion of the prediction injection hole corresponding to each of the prediction injection holes;
a prediction data acquisition unit that acquires prediction data of the ejection result of the prediction liquid ejection head from the learning model by inputting the prediction appearance image into a learning model obtained by performing machine learning using a learning appearance image corresponding to each of the learning ejection holes of a learning liquid ejection head that ejects liquid droplets from a plurality of learning ejection holes as an explanatory variable and a discharge result of the learning liquid ejection head as a target variable,
The appearance image acquisition unit acquires, as the prediction appearance image, an image of a region including at least a portion corresponding to the prediction injection hole and its surrounding portion in an actuator plate that controls the supply of liquid to the prediction injection hole.
Prediction system. The prediction system according to claim 1 , wherein the appearance image acquisition unit acquires, as the prediction appearance image, an image of the actuator plate at the time of assembly before being assembled into the prediction liquid jet head. The prediction system according to claim 1 or 2 , wherein the appearance image acquisition unit acquires a grayscale or colored image as the prediction appearance image. the ejection result as the objective variable is at least one of actual measurement data of an ejection angle, an ejection speed, and an amount of droplets, which is obtained by performing an ejection test on the learning liquid ejection head,
The prediction system according to claim 1 , wherein the prediction data obtained from the learning model is prediction data for at least one of the droplet ejection angle, ejection speed, and droplet volume that would be obtained if an ejection inspection were performed on the prediction liquid ejection head. The prediction system according to any one of claims 1 to 4, wherein the learning appearance images as explanatory variables include images of an injection hole plate in which the learning injection hole is formed before being incorporated into the learning liquid injection head, an actuator plate that controls the supply of liquid to the learning injection hole before being incorporated into the learning liquid injection head, and an image of either the injection hole plate or the actuator plate, which are bonded to each other , rotated a predetermined angle. The prediction system of any one of claims 1 to 4, wherein the training appearance images as explanatory variables include an image obtained by rotating, by a predetermined angle, an image of an injection hole plate in which the training injection hole is formed before being incorporated into the training liquid injection head, an actuator plate that controls the supply of liquid to the training injection hole before being incorporated into the training liquid injection head, and an image of either the injection hole plate or the actuator plate that are bonded to each other . A prediction method for predicting a discharge result for each of a plurality of prediction ejection holes of a prediction liquid ejection head that ejects liquid droplets from the plurality of prediction ejection holes, the method comprising the steps of:
acquiring a prediction appearance image including a peripheral portion of the prediction injection hole corresponding to each of the prediction injection holes;
a learning appearance image including a peripheral portion of each of the learning appearance images corresponding to each of the learning appearance images of a learning liquid ejection head that ejects liquid droplets from a plurality of learning ejection holes is used as an explanatory variable, and the ejection result of the learning liquid ejection head is used as an objective variable to perform machine learning, thereby inputting the prediction appearance image into a learning model obtained by inputting prediction data of the ejection result of the learning liquid ejection head from the learning model,
Further, as the prediction appearance image, an image of a region including at least a portion corresponding to the prediction injection hole and its surrounding portion in an actuator plate that controls the supply of liquid to the prediction injection hole is acquired.
including
Forecasting methods.

Images