JP2021159920A - Information processing device and program, information processing method, and information processing system - Google Patents

Abstract

To improve the efficiency of a blast process in a manufacturing process of a casting.SOLUTION: An information processing device (10) is provided with a controller (100). The controller (100) executes estimation processing for estimating blast conditions by using a learned model (M1) having undergone machine learning so as to derive output data related to blast conditions for executing a blast process from input data including pre-process information acquired before executing the blast process in a manufacturing process of a casting.SELECTED DRAWING: Figure 2

Description

The present invention relates to a technique for processing information related to a casting manufacturing process.

Techniques for processing information related to the manufacturing process of castings are known. For example, Patent Document 1 describes a technique for determining the dimensions of a hot water press, a weir, a runner, and the like used in the manufacturing process.

Japanese Unexamined Patent Publication No. 2000-326051 (published on November 28, 2000)

Here, in the manufacturing process of the cast product, the blasting process is executed as a part of the post-treatment process for the cast product taken out from the mold. The blasting step is a step of removing residual sand adhering to the casting product taken out from the mold. In the blasting process, a blasting device is used in which a projecting material is struck on the cast product to remove residual sand. The technique described in Patent Document 1 cannot improve the efficiency of the blasting process.

One aspect of the present invention is to realize a technique for streamlining a blasting process in a casting manufacturing process.

In order to solve the above problems, the information processing apparatus according to one aspect of the present invention includes a controller, and the controller includes input data including pre-process information obtained before the execution of the blast process in the manufacturing process of a casting product. From, the estimation process for estimating the blast condition is executed using the trained model machine-learned to derive the output data related to the blast condition for executing the blast process.

Here, the blasting conditions for executing the blasting step included in the manufacturing process of the cast product may differ depending on the information obtained from the steps prior to the blasting step. According to the above configuration, more appropriate blasting conditions can be estimated according to the pre-process information, so that the blasting process can be made more efficient.

In the information processing apparatus according to one aspect of the present invention, the blasting conditions include a blasting time for continuing the blasting process, a blasting power supplied in the blasting process, a projection amount of a projection material used in the blasting process, and the projection. It preferably comprises at least one of the projecting velocities of the material.

According to the above configuration, at least one of a more suitable blast time, blast power, projection amount, and projection speed can be estimated to streamline the blasting process.

In the information processing apparatus according to one aspect of the present invention, it is preferable that the input data further includes environmental information indicating an environment in which the manufacturing process is performed, in addition to the pre-process information.

Here, the blasting conditions for executing the blasting process included in the manufacturing process of the cast product may be further different depending on the environment in which the manufacturing process is performed. According to the above configuration, more appropriate blasting conditions can be estimated according to the environmental information, so that the blasting process can be further made more efficient.

In the information processing apparatus according to one aspect of the present invention, the controller applies the blast conditions estimated using the learned model in the estimation process to the recess information regarding the recess of the casting, the shape information regarding the shape, and the shape information regarding the shape. It is preferred to make corrections according to some or all of the surface information about the surface.

Here, the blasting conditions may differ depending on the recess information, the shape information, or the surface information. According to the above configuration, more appropriate blasting conditions can be estimated according to this information, so that the blasting process can be further made more efficient.

In the information processing apparatus according to one aspect of the present invention, the input data further includes a portion or all of recess information regarding the recess of the casting, shape information regarding the shape, and surface information regarding the surface of the casting. Is preferable.

Here, the blasting conditions may differ depending on the recess information, the shape information, or the surface information. Therefore, according to the above configuration, more appropriate blasting conditions can be estimated according to the recess information, the shape information, or the surface information, so that the blasting process can be further made more efficient.

In the information processing apparatus according to one aspect of the present invention, the controller is a machine that derives output data indicating the amount of residual sand in a casting taken out from a mold as output data related to the blast condition in the estimation process. It is preferable to estimate the blast condition based on the amount of residual sand indicated by the output data using the trained trained model.

Here, the blasting conditions may differ depending on the amount of residual sand. As an example, if the amount of remaining sand is large, a large amount of blasting time may be required. According to the above configuration, the blasting condition can be estimated according to the amount of residual sand, so that the blasting process can be further made more efficient.

In the information processing apparatus according to one aspect of the present invention, it is preferable that the controller further executes a generation process for generating the trained model by machine learning.

According to the above configuration, it is possible to generate a trained model used to achieve the above-mentioned effects.

In order to solve the above problems, the information processing apparatus according to one aspect of the present invention includes a controller, and the controller includes input data including pre-process information obtained before the execution of the blast process in the manufacturing process of a casting product. From, the generation process of generating the trained model for deriving the output data related to the blast condition for executing the blast process by machine learning is executed.

According to the above configuration, it is possible to generate a trained model used to achieve the above-mentioned effects.

In the information processing apparatus according to one aspect of the present invention, the controller uses learning data in which the pre-process information and the information related to the blast condition for executing the blast process are associated with each other in the generation process. , It is preferable to generate the trained model.

According to the above configuration, a trained model used to achieve the above-mentioned effects can be generated by supervised machine learning.

In order to solve the above problems, the program according to one aspect of the present invention is a program for controlling the above-mentioned information processing apparatus, and causes the controller to execute each of the above processes.

According to the above configuration, the same effect as that of the above-mentioned information processing apparatus is obtained.

In order to solve the above problems, the information processing method according to one aspect of the present invention is an information processing method executed by an information processing apparatus, and is obtained before the blasting process is executed in the casting manufacturing process. It includes an estimation step of estimating the blast condition using a trained model machine-learned to derive output data related to the blast condition for executing the blast step from input data including information.

According to the above configuration, the same effect as that of the above-mentioned information processing apparatus is obtained.

In order to solve the above problems, the information processing system according to one aspect of the present invention includes the above-mentioned information processing apparatus and a casting facility for carrying out the manufacturing process.

According to the above configuration, the same effect as that of the above-mentioned information processing apparatus is obtained.

According to one aspect of the present invention, the blasting process in the manufacturing process of a cast product can be made more efficient.

It is a block diagram which shows the structure of the information processing system which concerns on embodiment of this invention. It is a block diagram which shows the functional structure of the information processing apparatus which concerns on embodiment of this invention. It is a block diagram which shows an example of the hardware composition of the information processing apparatus which concerns on embodiment of this invention. It is a flowchart explaining the process flow in the information processing system which concerns on embodiment of this invention. It is a flowchart explaining the flow of the process which the information processing apparatus which concerns on embodiment of this invention generates a trained model. It is a figure explaining the trained model generated in embodiment of this invention. It is a flowchart explaining the flow of the process which estimates the blast condition by the information processing apparatus which concerns on embodiment of this invention. It is a figure explaining the trained model generated in the modification 1 of the Embodiment of this invention. It is a figure explaining the trained model and the estimation process in the modification 2 of the Embodiment of this invention. It is a figure explaining the trained model and the estimation process in the modification 3 of the embodiment of this invention.

[Embodiment]
Hereinafter, one embodiment of the present invention will be described in detail.

<Configuration of information processing device 10>
FIG. 1 is a block diagram showing a configuration of an information processing system 1 according to the present embodiment. As shown in FIG. 1, the information processing system 1 includes an information processing device 10 and a casting facility 90. The information processing device 10 and the casting equipment 90 are connected to each other via the network N1. For example, the information processing device 10 may be connected to each facility included in the casting facility 90 via a network N1. Further, for example, the information processing device 10 may be connected to a management device (not shown) that manages each facility included in the casting facility 90. The network N1 may be, for example, a wired LAN (Local Area Network), a wireless LAN, a mobile data communication network, the Internet, or a combination thereof.

<Overview of casting equipment 90>
The casting equipment 90 is an equipment for carrying out a manufacturing process of a cast product. The manufacturing process of the cast product includes, for example, a sand processing process, a molding process, a core molding process, a pouring process, a cooling transfer process, and a post-processing process. The post-treatment step includes a blast step. Details of each step will be described later. In the present embodiment, the cast product may indicate a cast product in which all the manufacturing processes have been completed, or may indicate a cast product after being taken out from the mold and before the post-treatment process is completed.

Further, the casting equipment 90 includes a molding equipment 91, a core molding equipment 92, a hot water pouring equipment 93, a cooling transfer equipment 94, a post-treatment equipment 95, and a sand treatment equipment 96.

The molding equipment 91 is equipment for carrying out a molding process of molding a mold from sand-treated mold sand.

The core molding equipment 92 is equipment for carrying out a core molding process for molding a core to be installed in a mold. The core molding equipment 92 may be provided in the molding equipment 91, or may be provided between the molding equipment 91 and the hot water pouring equipment 93.

The hot water pouring facility 93 is a facility for carrying out a hot water pouring process of producing molten metal and pouring it into a mold.

The cooling transport facility 94 is a facility that carries out a cooling transport step of transporting and cooling a molded mold, a mold in which molten metal is cast, a cast product taken out by disassembling the mold, and the like.

The post-treatment equipment 95 is equipment for carrying out a post-treatment step of performing post-treatment on the cast product taken out from the mold. The aftertreatment facility 95 includes a blasting device 951. The blasting apparatus 951 is an apparatus for carrying out a blasting step of removing residual sand adhering to a cast product taken out from a mold.

Here, one or more castings can be charged into the blasting apparatus 951. In other words, the blasting apparatus 951 can simultaneously perform the blasting process on one or more castings. Hereinafter, one or more castings in which the blasting apparatus 951 simultaneously performs the blasting step are also referred to as a processing unit. In this embodiment, for the sake of clarity of explanation, an example in which the number of cast products per processing unit is 1 will be described. An example in which the number of cast products per processing unit is plural will be described later as a modified example of the present embodiment.

The sand processing equipment 96 is equipment for processing mold sand used for molding a mold.

In addition, each of the above-mentioned equipment included in the casting equipment has a control unit (not shown). Each control unit transmits the data acquired in the equipment to the information processing device 10. The data transmitted by each control unit includes, for example, measured values and set values. The measured value is a value measured during the operation of each facility. The set value is a value set to operate each facility. For example, the set value is set to a value that suppresses the occurrence of defects in each equipment in each process.

<Functional configuration of information processing device 10>
Next, the functional configuration of the information processing device 10 will be described. FIG. 2 is a block diagram showing a functional configuration of the information processing device 10. As shown in FIG. 2, the information processing device 10 includes a controller 100. The controller 100 includes a learning phase execution unit 110, a storage unit 120, and an estimation phase execution unit 130. The learning phase execution unit 110 includes a generation unit 111. The storage unit 120 stores the learning data D1 and the learned model M1. The estimation phase execution unit 130 includes an input data acquisition unit 131 and an estimation unit 132.

<Example of hardware configuration of controller 100>
Next, an example of the hardware configuration of the controller 100 will be described. FIG. 3 is a diagram showing an example of the hardware configuration of the controller 100. As shown in FIG. 3, the controller 100 is composed of a computer including a processor 101, a memory 102, an input / output interface 103, and a communication interface 104.

The processor 101, the memory 102, the input / output interface 103, and the communication interface 104 are connected to each other via a bus. As the processor 101, for example, a CPU (Central Processing Unit), a microprocessor, a digital signal processor, a microcontroller, or a combination thereof is used. As the memory 102, for example, a semiconductor RAM (random access memory), an HDD (Hard Disk Drive), an SSD (Solid State Drive), or a combination thereof is used.

The memory 102 stores a program for causing the processor 101 to execute the generation method and the estimation method executed by the controller 100, which will be described later. The processor 101 executes the generation method and the estimation method by reading and executing the program stored in the memory 102. Further, the memory 102 stores various data referred to by the processor 101 in order to execute the generation method and the estimation method. Each of the generation method and the estimation method is an example of the information processing method in the present invention.

The input / output interface 103 is an interface for connecting an input / output device. For example, the input / output interface 103 is a USB (Universal Serial Bus) interface, a serial communication interface such as RS-232C, RS-422, RS-485, a short-range communication interface such as infrared rays and Bluetooth (registered trademark), or a short-range communication interface thereof. It may be composed of combinations.

An input device 105 and an output device 106 are connected to the input / output interface 103. As the input device 105, for example, a keyboard, a mouse, a touch pad, a microphone, or a combination thereof or the like is used. As the output device 106, for example, a display, a printer, a speaker, or a combination thereof is used.

The communication interface 104 is an interface that connects to the network N1. For example, the communication interface 104 may be an interface for performing wired LAN communication. In this case, the communication interface 104 is connected to a switching hub (not shown) and communicates with the casting equipment 90 via the switching hub. Further, the communication interface 104 may be an interface for performing wireless LAN communication. In this case, the communication interface 104 communicates with the casting equipment 90 via an access point (not shown).

<Process flow in information processing system 1>
FIG. 4 is a flowchart illustrating a flow of processing executed by the information processing system 1.

In step S1, the sand treatment equipment 96 executes the sand treatment step. In the sand treatment step, for example, the sand treatment facility 96 collects the mold sand from which the mold has been separated and the mold sand removed as residual sand from the casting. Further, the sand treatment equipment 96 executes processing for making the mold sand available in the molding equipment 91, such as removal of foreign substances, water injection / kneading, and the like. Further, the control unit included in the sand processing equipment 96 transmits the sand processing data acquired in the sand processing process to the information processing device 10.

In step S2, the molding equipment 91 executes the molding process. In the molding process, for example, the molding equipment 91 molds the upper mold and the lower mold. Further, the control unit included in the molding equipment 91 transmits the molding data acquired in the molding process to the information processing apparatus 10.

In step S3, the core molding equipment 92 executes the core molding step. In the core molding step, for example, the core molding equipment 92 installs a core between the separated upper mold and lower mold. After the core is installed, the upper mold and the lower mold are molded. After that, the molded mold is conveyed to the hot water pouring facility 93 by the cooling transfer facility 94. Further, the control unit included in the core molding equipment 92 transmits the core data acquired in the core molding process to the information processing apparatus 10.

In step S4, the pouring equipment 93 executes the pouring step. In the pouring step, for example, the pouring equipment 93 melts the material to produce a molten metal. Further, for example, the hot water pouring facility 93 pours the produced molten metal into the mold. Further, the control unit included in the hot water pouring equipment 93 transmits the hot water pouring data acquired in the hot water pouring process to the information processing device 10.

In step S5, the cooling transfer equipment 94 transfers the mold into which the molten metal is cast while cooling it. Further, the cooling transfer equipment 94 disassembles the mold and takes out the cast product. Further, the cooling transport facility 94 transports the taken-out cast product to the post-treatment facility 95 while cooling it. Further, the control unit included in the cooling transfer equipment 94 transmits the cooling transfer data acquired in the cooling transfer process to the information processing apparatus 10.

In step S6, the information processing apparatus 10 executes an estimation process for estimating the blast condition using the trained model M1. In the estimation process, the data transmitted from each equipment in each step of steps S1 to S5 is used. Here, the blast condition is a condition set when the blast step in the next step S7 is carried out. The details of the step and the details of the blast condition will be described later.

In step S7, the post-treatment equipment 95 executes the post-treatment step. In the blasting step included in the post-treatment step, the blasting apparatus 951 removes the residual sand adhering to the casting by striking the projecting material on the surface of the casting. In the blasting step, the blasting conditions estimated in step S6 are applied.

<Generation process executed by the information processing device 10>
Next, a generation process (process for executing the generation method) for generating the trained model M1 used in the estimation process in step S6 will be described. FIG. 5 is a flowchart illustrating a flow of generation processing executed by the information processing apparatus 10. The generation process shown in FIG. 5 is executed in advance before the processes of steps S1 to S7 shown in FIG. 4 are executed.

In step S101, the generation unit 111 generates the trained model M1 by machine learning. The trained model M1 was machine-learned to derive output data related to the blasting conditions for executing the blasting process from the input data including the pre-process information obtained before the blasting process is executed in the casting manufacturing process. It is a model. In the present embodiment, the input data further includes environmental information in addition to the pre-process information. Further, in the present embodiment, the output data related to the blast condition is information indicating the blast condition itself. Further, the generation unit 111 stores the generated trained model M1 in the storage unit 120.

At this time, the generation unit 111 generates the trained model M1 using the training data D1. The learning data D1 is data in which pre-process information and environmental information are associated with blast conditions. In other words, the generation unit 111 is supervised learning so that when the pre-process information and the environmental information included in the learning data D1 are input, the information indicating the blast condition itself associated with the pre-process information and the environmental information is output. Generates the trained model M1 by. Details of pre-process information, environmental information, and blast conditions will be described later.

A known technique can be applied to supervised learning of the trained model M1. For example, deep learning software (algorithm) such as CNN (Convolutional Neural Networks) or RNN (Recurrent Neural Network) can be used for supervised learning of the trained model M1. Further, not only the deep learning software but also other machine learning algorithms such as a support vector machine may be applied to the supervised learning of the trained model M1.

<Details of trained model M1>
FIG. 6 is a diagram for explaining the trained model M1 generated by the generation process. As shown in FIG. 6, when the trained model M1 acquires the input data including the pre-process information and the environmental information, it outputs the output data indicating the blast condition. As described above, the output data indicating the blast condition is an example of the "output data related to the blast condition" in the present invention.

(Pre-process information)
Here, the pre-process information is information obtained before the execution of the blasting process in the manufacturing process of the cast product. For example, pre-process information includes information obtained in one or more processes performed prior to the blasting process. In the present embodiment, the one or more steps performed before the blast step are a sand treatment step, a molding step, a core molding step, a pouring step, a cooling transfer step, and a post-treatment step. In this case, the pre-process information includes sand processing data, molding data, core data, pouring data, and cooling transfer data. Also, for example, the pre-process information includes the conditions specified in the execution of the blast process. The conditions specified in the execution of the blasting step include, for example, the type of projecting material.

Specifically, the sand treatment data includes the temperature, moisture content, CB value, air permeability, air pressure, and compressive strength of the mold sand.

The molding data includes the upper frame board set complete position, upper frame mold thickness (mold thickness), upper frame compression rate (mold compression rate), lower frame board set complete position, lower frame mold thickness (mold thickness), and lower frame. Compression rate (mold compression rate), squeeze pressure (lower frame squeeze pressure peak, upper frame squeeze pressure peak), aeration pressure and sand tank pressure when aeration is completed, aeration pressure and sand tank pressure when exhaust is completed, upper frame mold thickness Target position, upper frame board set position, lower frame mold thickness target position, lower frame board set position, pattern thickness, number of times of waiting for upper type spray in frame, time of upper type spray in frame, number of times of waiting for lower type spray in frame, in frame Lower spray time, aeration pattern, chiller set execution, gas hole drilling, turntable speed, match plate type, sprue number, aeration upper frame board operating time, aeration upper frame board operating position, aeration Includes lower frame board operating time, lower frame board operating position during aeration, pouring pattern number, number of times to wait for out-of-frame spray, and part or all of out-of-frame spray time.

In addition, the pouring data includes the casting weight, the casting time, the amount of the inoculant added, the elapsed time of receiving the hot water, the casting temperature, and a part or all of the amount of the hot water discharged.

In addition, the cooling transfer data includes the cooling time. The cooling time indicates the length of time that the casting taken out of the mold is cooled.

The type of projection material is the type of projection material used in the blasting process. In this embodiment, there are a plurality of types of projection materials that can be used in the blasting process. The size of the projection material varies depending on the type. The blasting conditions may differ depending on the type of projecting material.

(Environmental information)
Here, the environmental information is information indicating the environment in which the manufacturing process of the cast product is carried out. For example, environmental information includes some or all of the air temperature, humidity, and weather in the space where the casting equipment 90 is installed.

(Blast condition)
The blast condition is a condition set when the blast process is carried out. In the present embodiment, the blasting conditions include the blasting time, the blasting power, the projection amount of the projection material, and the projection speed of the projection material.

The blast time is the time to continue the blasting process. If the blasting time is too short, the residual sand adhering to the cast product taken out from the mold cannot be sufficiently removed, but if the blasting time is too long, power consumption, required time, and the like are wasted. Therefore, it is desirable to set an appropriate blast time. The appropriate blasting time varies depending on, for example, the amount of sand remaining on the cast product taken out from the mold, the shape of the cast product, and the like.

The blast power is the power supplied to the blast device 951 in the blast process. By properly setting the blast power, the power can be consumed efficiently. The appropriate blast power varies depending on the projection amount of the projection material, the type of the projection material, the projection speed of the projection material, and the like.

The projection amount of the projection material is the amount of the projection material projected in the blasting process. For example, the projection amount may be the total amount of the projection material projected from the start to the end of the blasting process of one processing unit, but in the present embodiment, the amount of the projection material projected per unit time. Suppose that The appropriate projection amount depends on, for example, the shape of the cast product, the amount of residual sand adhering to the cast product taken out from the mold, and the like.

The projection speed of the projection material is the speed at which the projection material is projected in the blasting process. The appropriate projection speed depends on, for example, the shape of the cast product, the amount of residual sand adhering to the cast product taken out from the mold, and the like.

<Estimation processing executed by the information processing device 10>
Next, the details of the estimation process in step S6 will be described. FIG. 7 is a flowchart illustrating the flow of the estimation process (process of executing the estimation method) executed by the information processing apparatus 10.

In step S201, the input data acquisition unit 131 acquires pre-process information and environmental information as input data.

Specifically, the input data acquisition unit 131 receives data (sand processing data, molding data, core data) transmitted from each facility to the information processing apparatus 10 in each process of steps S1 to S5 as pre-process information. , Pouring data, and cooling transfer data), and the conditions specified in the execution of the blasting process (here, the type of projecting material) are acquired. Further, the input data acquisition unit 131 acquires environmental information as environmental information, for example, from a sensor (not shown) that detects the environment. The input data acquisition unit 131 may acquire a part or all of the input data via the input device 105.

In step S202, the estimation unit 132 estimates the blast condition using the learned model M1 stored in the storage unit 120. Specifically, the estimation unit 132 outputs the blast condition indicated by the output data to the output device 106 as the estimated blast condition.

<Effect of this embodiment>
In the manufacturing process of the cast product, the information processing system 1 according to the present embodiment estimates and estimates the blasting conditions by referring to the pre-process information obtained before the execution of the blasting process and the environmental information in which the manufacturing process is performed. The blasting process is performed using the blasting conditions. Therefore, the blasting process can be carried out using blasting conditions suitable for factors that may differ for each manufacturing process, and the blasting process can be made more efficient.

[Modification 1]
The information processing system 1 according to the present embodiment can be modified to use the trained model M2 instead of the trained model M1.

<Details of trained model M2>
FIG. 8 is a diagram illustrating input / output data of the trained model M2. As shown in FIG. 6, when the trained model M2 acquires the input data including the pre-process information, the environmental information, the concave portion information, the shape information, and the surface information, the trained model M2 outputs the output data indicating the blast condition. In other words, the input data input to the trained model M2 includes recess information, shape information, and surface information in addition to the data input to the trained model M1. The output data output from the trained model M2 is the same as the output data output from the trained model M1.

(Recess information)
Here, the recess information is information indicating the recess of the cast product. The recesses include those formed on the surface of the casting and those formed on the inner surface forming the boundary between the casting and the internal space. The internal space is the space formed by the installation of the core. The recess on the surface or the inner surface is a portion that satisfies a predetermined condition. The portion satisfying the predetermined condition may be, for example, a portion having an opening and a depth. Further, the portion satisfying the predetermined condition may be a portion sandwiched between the surface of the convex portion and the surface connected to the root of the convex portion. The recess can be extracted as a portion satisfying a predetermined condition from the shape information described later.

The amount of sand remaining in the casting taken out of the mold may vary depending on the degree of recess indicated by the recess information. As an example, the larger the degree of the recess, the larger the amount of residual sand tends to be. The degree of recesses can vary, for example, depending on the number and shape of the recesses. The shape of the recess includes the depth, the opening area, the opening shape, and the like. For example, in the case of recesses having the same shape, the greater the number of recesses, the greater the degree of recesses. Further, the deeper the recess, the larger the degree of the recess. Further, the smaller the opening area of the recess, the larger the degree of the recess. Further, the larger the aspect ratio of the opening shape of the recess (that is, the longer the recess), the larger the degree of the recess. For example, the recess information may be represented by one of the above-mentioned number, depth, opening area, opening shape, and the like. Further, the recess information may be represented by a combination of a plurality of parameters as described above. In this case, the recess information may be information obtained by weighting and combining each parameter.

Further, it is considered that the degree of the recess is larger when there is one deep recess than when there are a plurality of shallow recesses. In this case, the information indicating the recess may be represented by the depth of the deepest recess among the one or more recesses.

For example, the recess information can be obtained from the pattern data of the casting. Further, for example, the recess information can be detected from an image obtained by capturing the mold with an imaging device (for example, a camera or the like). Further, for example, the concave portion information can be detected by using a sensor (for example, a three-dimensional scanner or the like) that detects the three-dimensional shape of the mold. The recess information may be detected by the control unit of the molding equipment 91 in the molding process and may be included in the molding data which is one of the pre-process information.

(Shape information)
The amount of sand remaining in the casting taken out of the mold may vary depending on the shape of the casting. As an example, the more complicated the shape, the larger the amount of sand remaining, and conversely, the smaller the surface area, the smaller the amount of sand remaining. The shape information is information on the shape of the casting. The shape information may include, for example, parameters that define the shape, may include three-dimensional shape data, or may include information indicating the surface area. The parameters that define the shape are, for example, the radius and height of the bottom surface of the cylinder, but are not limited to this. The three-dimensional shape data may be pattern data of the casting, or may be data detected from the casting immediately before the blasting step by using a sensor. For example, the three-dimensional shape data may be obtained by analyzing a two-dimensional image generated by a camera as a sensor. Further, for example, the three-dimensional shape data may be obtained by a three-dimensional shape sensor that detects the three-dimensional shape.

(Surface information)
The amount of sand remaining in the casting taken out of the mold may vary depending on the properties of the casting. As an example, the more the surface of the cast product is made of a material to which casting sand is less likely to adhere (the sand falls better), the smaller the amount of residual sand tends to be. The surface information is information about the surface of the casting. The surface information may include, for example, information for identifying the material constituting the surface of the casting, information indicating the characteristics of the surface, or information indicating the roughness of the surface. May be good.

<Generation process in this modified example>
In this modification, in step S101 shown in FIG. 5, the generation unit 111 uses the learning data D2 instead of the learning data D1 to generate the trained model M2. The learning data D2 is information in which pre-process information, environmental information, recess information, shape information, and surface information are associated with blast conditions.

<Estimation processing in this modified example>
In this modification, in step S201 shown in FIG. 7, the input data acquisition unit 131 acquires pre-process information, environmental information, recess information, shape information, and surface information as input data.

For example, the recess information, the shape information, and a part or all of the surface information may be included in the molding data transmitted by the control unit of the molding equipment 91 in step S2. In this case, the control unit of the molding equipment 91 acquires a part or all of the concave portion information, the shape information, and the surface information from the pattern data obtained in advance, includes it in the molding data, and transmits it. These pieces of information contained in the molding data indicate pre-designed design values for the shape of the casting.

Further, a part or all of the concave portion information, the shape information, and the surface information may be included in the cooling transfer data transmitted by the control unit of the cooling transfer equipment 94 in step S5. In this case, in step S5, the control unit of the cooling transfer equipment 94 acquires a part or all of the recess information, the shape information, and the surface information by using an imaging device (for example, a camera, a three-dimensional shape scanner, etc.) and cools the equipment. Include it in the transport data and send it. These pieces of information contained in the cooling transfer data indicate actual measurements of the shape of the casting taken out of the mold.

For example, the input data acquisition unit 131 may use the information included in the cooling transfer data when a part or all of the recess information, the shape information, and the surface information is not included in the molding data. In other words, the input data acquisition unit 131 may use the actually measured value when the design value indicating the shape of the cast product cannot be obtained. Further, the input data acquisition unit 131 may use information included in both the molding data and the cooling transfer data for a part or all of the concave portion information, the shape information, and the surface information. In other words, the input data acquisition unit 131 may use both the design value and the actually measured value indicating the shape of the cast product. This is effective when the difference between the design value and the measured value is large.

Since the process of step S202 is as described with reference to FIG. 7, the detailed description is not repeated.

<Effect of variant 1>
In the manufacturing process of the cast product, the information processing system 1 according to this modification refers to a part or all of the concave portion information, the shape information, and the surface information of the cast product in addition to the pre-process information and the environmental information. Estimate the blast condition. Therefore, the blasting process can be carried out using more appropriate blasting conditions according to the concave portion information, the shape information, and the surface information in which the amount of residual sand can be a different factor, and the blasting process can be further made more efficient. ..

[Modification 2]
The information processing system 1 according to the present embodiment can modify the estimation process as follows.

<Estimation processing in this modified example>
FIG. 9 is a schematic diagram illustrating an estimation process in this modified example. In this modification, in step S202, the estimation unit 132 corrects the blast condition estimated using the learned model M1 according to a part or all of the recess information, the shape information, and the surface information. The details of the recess information, the shape information, and the surface information are as described above.

For example, the estimation unit 132 corrects the blast time output from the trained model M1 so that the greater the degree of the recess indicated by the recess information, the longer the blast time. Further, the estimation unit 132 corrects the blast time output from the trained model M1 so that the more complicated the shape indicated by the shape information, the longer the blast time. Further, the estimation unit 132 corrects the blast time output from the trained model M1 so that the blast time becomes longer as the residual sand easily adheres to the material constituting the surface. Further, the estimation unit 132 corrects the blast time output from the trained model M1 so that the larger the surface area, the longer the blast time.

Further, for example, the estimation unit 132 corrects the blast power output from the learned model M1 so as to increase as the degree of the recess indicated by the recess information increases. Further, the estimation unit 132 corrects the blast power output from the learned model M1 so as to increase as the shape indicated by the shape information becomes more complicated. Further, the estimation unit 132 corrects the blast power output from the trained model M1 so as to increase so that residual sand easily adheres to the material constituting the surface. Further, the estimation unit 132 corrects the blast power output from the trained model M1 so that the larger the surface area, the larger the blast power.

Further, for example, the estimation unit 132 corrects the projection amount per unit time output from the learned model M1 so as to increase as the degree of the recess indicated by the recess information increases. Further, the estimation unit 132 corrects the projection amount per unit time output from the learned model M1 so as to increase as the shape indicated by the shape information becomes more complicated. Further, the estimation unit 132 corrects the projection amount per unit time output from the trained model M1 so as to increase so that residual sand easily adheres to the material constituting the surface. Further, the estimation unit 132 corrects the projection amount per unit time output from the trained model M1 so that the larger the surface area, the larger the projection amount.

Further, for example, the estimation unit 132 corrects the projection speed output from the learned model M1 so as to increase as the degree of the recess indicated by the recess information increases. Further, the estimation unit 132 corrects the projection speed output from the trained model M1 so as to increase as the shape indicated by the shape information becomes more complicated. Further, the estimation unit 132 corrects the projection speed output from the trained model M1 so as to increase so that residual sand easily adheres to the material constituting the surface. Further, the estimation unit 132 corrects the projection speed output from the trained model M1 so that the larger the surface area, the higher the projection speed.

<Effect of variant 2>
The information processing system 1 according to this modification corrects the blast condition estimated by referring to the pre-process information and the environmental information in the manufacturing process of the cast product by using the concave portion information of the cast product. Therefore, the blasting process can be performed using more appropriate blasting conditions according to the concave portion information, the shape information, and the surface information in which the amount of residual sand can be a different factor, and the blasting process can be further made more efficient. can.

[Modification 3]
The information processing system 1 according to the present embodiment can use the trained model M3 instead of the trained model M1 and can modify the estimation process as described later.

<Details of trained model M3>
FIG. 10 is a diagram illustrating an estimation process using the trained model M3. As shown in FIG. 10, when the trained model M3 acquires input data including pre-process information, environmental information, recess information, shape information, and surface information, it outputs output data indicating the amount of residual sand. In the present embodiment, the blast condition can be calculated from the output data indicating the amount of residual sand by a predetermined calculation process. The output data indicating the amount of residual sand is an example of "information related to the blast condition" in the present invention.

In other words, the output data output from the trained model M3 indicates the amount of residual sand and is different from the output data output from the trained models M1 and M2. The input data input to the trained model M3 is the same as the input data input to the trained model M2. However, the input data input to the trained model M3 may be the same as the input data input to the trained model M1.

<Generation process in this modified example>
In this modification, in step S101 shown in FIG. 5, the generation unit 111 uses the learning data D3 instead of the learning data D1 to generate the trained model M3 described above. The learning data D3 is information in which pre-process information, environmental information, recess information, shape information, and surface information are associated with information indicating the amount of residual sand.

<Estimation processing in this modified example>
In this modification, in step S202, the estimation unit 132 estimates the blast condition based on the amount of residual sand estimated using the trained model M3. For example, the estimation unit 132 calculates the blast condition, the blast power, the projection amount per unit time, and the projection speed by performing a predetermined arithmetic process on the residual sand amount indicated by the output data from the trained model M3.

<Effect of variant 3>
The information processing system 1 according to this modification estimates the blast condition based on the amount of residual sand estimated by referring to the pre-process information, the environmental information, the concave portion information, the shape information, and the surface information in the manufacturing process of the cast product. do. Therefore, the blasting step can be carried out using appropriate blasting conditions according to the estimated amount of residual sand, and the blasting step can be further made more efficient.

<Modification example 4>
The information processing system 1 according to the present embodiment can modify the processing in step S202 as follows when the blasting process is performed on a plurality of cast products in the same processing unit.

<Processing in step S202>
In this modification, in step S202, the estimation unit 132 acquires information on a plurality of cast products that simultaneously execute the blasting step in the processing unit. Information about the plurality of cast products includes, for example, the number of cast products, surface area information of each cast product, and the like.

Next, the estimation unit 132 performs an operation to convert the blast condition output from the trained model M1 into the blast condition corresponding to the plurality of target castings.

For example, the estimation unit 132 estimates the output values of the blast time, blast power, and projection speed of the output data from the trained model M1 as they are as blast conditions corresponding to a plurality of cast products.

Further, for example, the estimation unit 132 calculates the projection amount per unit area from the projection amount per unit time in the output data from the trained model M1. Further, the estimation unit 132 uses the projection amount per unit area to convert the projection amount corresponding to the total surface area of the plurality of cast products, and estimates the projection amount per unit time corresponding to the plurality of cast products. ..

The arithmetic processing for converting the blast condition output from the trained model M1 into the blast condition corresponding to the plurality of target castings is not limited to the above-mentioned example. For example, among the blast conditions output from the trained model M1, the conditions that can be applied as blast conditions corresponding to a plurality of cast products as they are are not limited to the above-mentioned examples. Further, the arithmetic processing applied to the condition requiring conversion is not limited to the sum described above, and weighted sum, average, weighted average, maximum value, minimum value, other arithmetic, or a combination thereof can be applied.

<Effect of variant 4>
The information processing system 1 according to this modification can carry out the blasting process using more appropriate blasting conditions when a plurality of castings are simultaneously put into the blasting apparatus 951 in the casting product manufacturing process. The blasting process can be further made more efficient.

[Other variants]
In the above-described embodiment of the present invention, an example in which the generation unit 111 generates the trained model M1 by supervised machine learning has been described. Not limited to this, the generation unit 111 may generate the trained model M1 by unsupervised machine learning. For example, when estimating a stepwise blast power as a blast condition, it is possible to generate a trained model M1 by unsupervised machine learning and associate the clustering result output from the trained model M1 with the stage of the blast power. It is possible. For example, deep learning software (algorithm) such as GAN (Generative Adversarial Network) can be applied to unsupervised machine learning. The learned model M1 is not limited to supervised learning and unsupervised learning, and may be machine-learned by semi-supervised learning or the like.

Further, in the present embodiment, the output data output from the trained model M1 may further include the type of projection material. In this case, the trained model M1 may output the same type as the type of projection material included in the input data.

Further, in the present embodiment, the input data input to the trained model M1 has been described as including the data obtained in all the processes performed before the blasting process as the pre-process information. Not limited to this, the input data may include data obtained in at least one step performed before the blast step.

Further, in the present embodiment, the input data input to the trained model M1 has been described as including the environmental information, but it does not necessarily have to include the environmental information.

Further, in the present embodiment, the information processing apparatus 10 may use a trained model in which information including the amount of residual sand is used as input data. In this case, the generation unit 111 generates a trained model by machine learning using the learning data in which the amount of remaining sand and the blast condition are associated with each other. Further, the estimation unit 132 calculates the amount of residual sand by arithmetic processing based on one or both of the pre-process information and the environmental information. Further, the estimation unit 132 estimates the blast condition by inputting the amount of residual sand calculated by the arithmetic processing into the trained model.

Further, in the present embodiment, the information processing device 10 does not have to use the trained model. In this case, the information processing device 10 does not have to include the generation unit 111. The estimation unit 132 calculates the amount of residual sand by arithmetic processing based on one or both of the pre-process information and the environmental information. Further, the estimation unit 132 calculates the blast condition by the calculation process based on the amount of remaining sand calculated by the calculation process.

Further, in the present embodiment, an example in which the information processing apparatus 10 includes the learning phase execution unit 110 and the estimation phase execution unit 130 has been described. The functions realized by these components may be distributed and realized in a plurality of devices. For example, the information processing device 10 may be composed of a learning device and an estimation device that is physically different from the learning device. In this case, the learning device includes the learning phase execution unit 110, the estimation device includes the estimation phase execution unit 130, and the determination device executes the estimation process using the learned model M1 generated by the learning device. Such a learning device and a determination device are each included in the category of the present invention.

Further, in the present embodiment, the information processing device 10 has been described as having a storage unit 120 for storing the learned model M1. Not limited to this, at least one of the training data D1 and the trained model M1 may be stored in an external device. In this case, at least one of the learning phase execution unit 110 and the estimation phase execution unit 130 performs processing with reference to the learning data D1 or the learned model M1 stored in the external device.

Further, in the present embodiment, the control block of the information processing apparatus 10 (particularly, the generation unit 111, the input data acquisition unit 131, and the estimation unit 132) has been described as being realized by software. Not limited to this, these control blocks may be realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like.

When these control blocks are realized by software, a program that realizes an example of the information processing method in the present invention has been described as being stored in the memory 102. However, the recording medium for recording the above program is not limited to the memory 102, and is not limited to the memory 102, but is a "non-temporary tangible medium" such as a ROM (Read Only Memory), a tape, a disk, a card, a semiconductor memory, or a programmable medium. A logic circuit or the like can be used. Further, a RAM (Random Access Memory) for expanding the above program may be further provided. Further, the program may be supplied to the computer via an arbitrary transmission medium (communication network, broadcast wave, etc.) capable of transmitting the program. It should be noted that one aspect of the present invention can also be realized in the form of a data signal embedded in a carrier wave, in which the above program is embodied by electronic transmission.

[Additional notes]
The present invention is not limited to the above-described embodiments, and various modifications can be made within the scope of the claims, and the embodiments obtained by appropriately combining the technical means disclosed in the different embodiments. Is also included in the technical scope of the present invention.

1 Information processing system 10 Information processing equipment 90 Casting equipment 91 Molding equipment 92 Core molding equipment 93 Pouring equipment 94 Cooling transfer equipment 95 Post-processing equipment 951 Blasting equipment 96 Sand processing equipment 100 Controller 101 Processor 102 Memory 103 Input / output interface 104 Communication Interface 105 Input device 106 Output device 110 Learning phase execution unit 111 Generation unit 120 Storage unit 130 Estimating phase execution unit 131 Input data acquisition unit 132 Estimating unit

Claims (12)

Equipped with a controller
The controller
Using a trained model machine-learned to derive output data related to the blasting conditions for executing the blasting process from input data including pre-process information obtained before the blasting process is executed in the casting manufacturing process. Then, the estimation process for estimating the blast condition is executed.
An information processing device characterized by this. The blasting condition includes at least one of a blasting time for continuing the blasting process, a blasting power supplied in the blasting process, an amount of projecting material used in the blasting process, and a projecting speed of the projecting material.
The information processing apparatus according to claim 1. The input data further includes environmental information indicating an environment in which the manufacturing process is performed, in addition to the pre-process information.
The information processing apparatus according to claim 1 or 2. In the estimation process, the controller makes the blast condition estimated using the trained model a part or all of the recess information regarding the concave portion of the casting, the shape information regarding the shape, and the surface information regarding the surface. To correct,
The information processing apparatus according to any one of claims 1 to 3, wherein the information processing device is characterized by the above. The input data further includes some or all of the recess information about the recesses of the casting, the shape information about the shape, and the surface information about the surface of the casting.
The information processing apparatus according to any one of claims 1 to 3, wherein the information processing device is characterized by the above. The controller uses a trained model machine-learned to derive output data indicating the amount of sand remaining in the cast product taken out from the mold as output data related to the blast condition in the estimation process. The blast condition is estimated based on the amount of residual sand indicated by the data.
The information processing apparatus according to any one of claims 1 to 5, wherein the information processing device is characterized by the above. The controller
Generation processing to generate the trained model by machine learning,
To do more,
The information processing apparatus according to any one of claims 1 to 6, wherein the information processing device is characterized by the above. Equipped with a controller
The controller
Generation by machine learning to derive a trained model that derives output data related to the blasting conditions for executing the blasting process from input data including pre-process information obtained before the blasting process is executed in the casting manufacturing process. Process, execute,
An information processing device characterized by this. The controller
In the generation process, the trained model is generated using the learning data in which the pre-process information and the information related to the blast condition for executing the blast process are associated with each other.
The information processing apparatus according to claim 8. A program for controlling the information processing apparatus according to any one of claims 1 to 9, wherein the controller executes each of the processes. It is an information processing method executed by an information processing device.
Using a trained model machine-learned to derive output data related to the blasting conditions for executing the blasting process from input data including pre-process information obtained before the blasting process is executed in the casting manufacturing process. The estimation process for estimating the blast condition,
Information processing methods including. The information processing device according to any one of claims 1 to 9.
The casting equipment that carries out the manufacturing process and
Information processing system including.

Images