Automatic system and method for determining whether a shot is a discarded shot.

Abstract

To provide a preheating shot automatic determination system and a preheating shot automatic determination method which can automatically execute switching determination from preheating shot casting to production casting at high precision with high efficiency using preheating shot waveform data which can be confirmed altogether with respect to many factors which are regarded to affect the casting quality of a cast product.SOLUTION: Provided are a preheating shot automatic determination system and a preheating shot automatic determination method having an automatic determination part and an automatic determination step which can automatically execute switching determination using preheating shot waveform data on preheating shot casting upon executing switching determination from preheating shot casting in which the injection filling of molten metal toward a mold cavity is repeated based on preset preheating shot conditions to production casting in which the molten metal is injection-filled into the mold cavity based on preset production conditions to cast a produced cast product.SELECTED DRAWING: Figure 1

Description

Technical Field

[0003]

[0001] The present invention relates to a throw-away automatic determination system and a throw-away automatic determination method for performing a switching determination from a throw-away casting that repeatedly injects and fills molten metal into a mold cavity based on preset throw-away conditions to a production casting that injects and fills molten metal into a mold cavity based on preset production conditions to manufacture a production casting product.

Background Art

[0002] In casting molding in which molten metal such as an aluminum alloy supplied to an injection part is injected and filled into a mold cavity formed by clamping a casting mold, for the purpose of stabilizing the quality of the casting product and adjusting the casting conditions, etc., before the production casting for manufacturing the production casting product, it is common to perform throw-away casting. This throw-away casting is a process for preheating the casting mold and the injection part to a temperature suitable for casting molding by utilizing the heat of the high-temperature molten metal. Also, for example, in the case of an injection part driven by hydraulic pressure, it is a process for preheating the temperature of the hydraulic oil to a suitable state. Further, in order to adjust the casting conditions, for example, it is a process for checking the flow state of the molten metal in the mold cavity, and a process for checking the pressure applied to the molten metal suitable for the molten metal to cool and solidify in the mold cavity preheated to a suitable temperature.

[0003] <Production casting begins after this initial test casting. Therefore, it is highly desirable to accurately preheat the casting mold, injection unit, and hydraulic fluid to suitable temperatures during the test casting, and to appropriately adjust the casting conditions. Furthermore, from the standpoint of production efficiency in casting, it is preferable to minimize the number of test casting steps as much as possible. In contrast, the number of test castings was determined based on past casting performance, and the decision to switch from test casting to production casting was based on this set number. Alternatively, the switchover decision was made by visual inspection by the casting technician. However, this conventional switchover decision method is ambiguous in determining whether the original purpose of the test casting has been achieved, often resulting in an excessive number of casts being set to allow for a margin of safety, thus failing to improve production efficiency. Therefore, many proposals have been made to enable accurate determination of the switchover from test casting to production casting.

[0004] For example, a method has been proposed in which a temperature measuring device, such as the one shown in Patent Document 1, is used to capture thermal images and calculate the temperature of the mold, thereby determining the switch from sacrificial casting to production casting based on the mold temperature. According to this method, the temperature distribution of the mold cavity due to preheating of the casting mold by molten metal and cooling of the casting mold by cooling water can be accurately measured, enabling stable continuous molding in production casting. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2021-41435 [Overview of the project] [Problems that the invention aims to solve]

[0006] Here, the method described in Patent Document 1 states that the decision to switch from test casting to production casting is made solely by measuring the temperature of the casting mold. Furthermore, if interpreted more broadly, it is also possible to measure the temperature of the injection section and hydraulic fluid using a temperature measuring device, and it is conceivable that the decision to switch could be made using these temperature measurements. However, many factors other than temperature are intricately intertwined and affect the casting quality of casting.

[0007] For example, in a casting apparatus (called a die-casting machine) equipped with an injection sleeve and a plunger tip, irregular thermal deformation occurs due to the heat of the molten metal, causing it to bend significantly (known as banana deformation), increasing the sliding resistance between the injection sleeve and the plunger tip, disrupting the behavior of injection filling of the molten metal into the mold cavity, and potentially causing significant fluctuations in casting quality. Alternatively, the behavior of injection filling can also be disrupted by galling, where molten metal penetrates the sliding surface between the injection sleeve and the plunger tip. Furthermore, many other factors are involved, such as fluctuations in the amount of molten metal supplied to the injection sleeve, the influence of molten metal residues such as oxides and impurities contained in the molten metal, fluctuations in the composition of the molten metal, aging of the plunger tip's driving mechanism, mold residue defects where a portion of the casting remains in the mold cavity, and the state of release agent application to the mold cavity. These factors cannot be detected by temperature measurement alone, and the means shown in Patent Document 1, including its broader interpretation, are not suitable as means for accurately determining when to switch from test casting to production casting.

[0008] Therefore, the present invention aims to provide an automatic discard casting determination system and method that can automatically and efficiently determine when to switch from discard casting to production casting with high accuracy, using discard casting waveform data that can be used to check for many factors that are said to affect the casting quality of castings all at once. [Means for solving the problem]

[0009] The automatic discard detection system of the present invention In an automated sacrificial casting determination system, which determines whether to switch from sacrificial casting, in which molten metal is repeatedly injected and filled into the mold cavity based on pre-set sacrificial casting conditions, to production casting, in which molten metal is injected and filled into the mold cavity based on pre-set production conditions to cast production products, The system is characterized by having an automatic determination unit that automatically performs the switching determination using the waveform data of the aforementioned test casting.

[0010] In the automatic determination system of the present invention, The automatic determination unit preferably includes an AI creation mode and an AI determination mode.

[0011] Furthermore, in the automatic determination system of the present invention, The AI ​​creation mode preferably includes a data transmission / reception unit that receives the discard waveform data and transmits the result of the switching determination; a data setting unit that selects waveform data to be used for the switching determination from the discard waveform data and sets the determination conditions; and a learning unit that creates and saves an AI learning model in which the learning interval of the waveform data has been set and the features have been defined.

[0012] Furthermore, in the automatic determination system of the present invention, Preferably, the AI ​​judgment mode includes a judgment unit that performs the switching judgment by comparing the AI ​​learning model with the discard waveform data, and a display unit that displays the progress of the switching judgment and the judgment result.

[0013] The automatic determination method of the present invention is In an automatic discard casting determination method, which determines whether to switch from a discard casting process, in which molten metal is repeatedly injected and filled into the mold cavity based on pre-set discard casting conditions, to a production casting process, in which molten metal is injected and filled into the mold cavity to cast production products based on pre-set production conditions, The invention is characterized by comprising an automatic determination process that automatically performs the switching determination using the waveform data of the test casting process.

[0014] In the automatic determination method of the present invention, The aforementioned automatic determination process preferably comprises an AI creation process and an AI determination process.

[0015] Furthermore, in the automatic determination method of the present invention, The AI ​​creation process preferably includes a waveform data collection process that receives the discard waveform data and transmits the result of the switching determination; a data setting process that selects waveform data to be used for the switching determination from the discard waveform data and sets the determination conditions; and a learning model creation process that creates an AI learning model in which the learning interval of the waveform data is set and the features are defined.

[0016] Furthermore, in the automatic determination method of the present invention, Preferably, the AI ​​determination step includes a training data setting step of setting the AI ​​learning model as training data, a switching determination step of performing the switching determination using a supervised learning program, and a display step of displaying the progress of the switching determination and the determination result.

[0017] Furthermore, in the automatic determination method of the present invention, Preferably, the switching determination step involves using a supervised learning program of a regression method to count the difference between the training data and the discarded waveform data as a determination score value, and then performing the switching determination by comparing the regression analysis value obtained by regression analysis of the determination score value with a preset threshold.

[0018] Furthermore, in the automatic determination method of the present invention, Preferably, the switching determination step involves using a supervised learning program for the classification method to compare the training data with the discarded waveform data and perform the switching determination. [Effects of the Invention]

[0019] According to the present invention, it is possible to provide a drop-forging automatic determination system and a drop-forging automatic determination method that can efficiently and automatically perform high-precision determination of the switching from drop-forging to production casting using drop-forging waveform data that can collectively confirm many factors that are considered to affect the casting quality of cast products.

Brief Description of the Drawings

[0020] [Figure 1] It is a conceptual diagram showing a drop-forging automatic determination system according to an embodiment of the present invention. [Figure 2] It is a diagram showing the automatic determination unit of the drop-forging automatic determination system shown in FIG. 1. [Figure 3] It is a diagram showing a drop-forging automatic determination method according to an embodiment of the present invention. <( [Figure 4] It is a diagram showing the drop-forging casting waveform data of the drop-forging automatic determination method shown in FIG. 3. [Figure 5] It shows a form in which other drop-forging casting waveform data is selected.

Modes for Carrying Out the Invention

[0021] Hereinafter, preferred embodiments for carrying out the present invention will be described with reference to the drawings. Note that the following embodiments do not limit the invention according to each claim. Also, not all combinations of the features described in the embodiments are essential for the solution means of the invention according to each claim. Further, in the present embodiment, the scales and dimensions of each component may be shown exaggerated, or some components may be omitted. [[ID=3)]]

[0022] (Drop-forging Automatic Determination System) First, the drop-forging automatic determination system according to an embodiment of the present invention will be described with reference to FIG. 1. The drop-forging automatic determination system 200 shown in FIG. 1 is configured to connect an automatic determination unit 50 to a casting device 100. <(

[0023] The casting apparatus 100 comprises a casting mold 10, a clamping section 20, an injection section 30, and a casting control section 40. In Figure 1, the casting mold 10 and the injection section 30 are shown arranged horizontally (horizontal clamping horizontal casting apparatus 100), but the apparatus is not limited to this configuration. For example, the casting mold 10 may be arranged horizontally and the injection section 30 may be arranged vertically (horizontal clamping vertical casting apparatus 100), or the casting mold 10 and the injection section 30 may be arranged vertically (vertical clamping vertical casting apparatus 100). In any case, the components remain the same, only the combination of arrangement of the casting mold 10 and the injection section 30 is changed, so the explanation will use the horizontal clamping horizontal casting apparatus 100.

[0024] The casting mold 10 forms a mold cavity 13 and a mold gate 14 by operating the clamping section 20 to clamp the fixed platen 21 and the movable platen 22. It is preferable to apply a release agent to the mold cavity 13 and the mold gate 14 before injecting and filling the molten metal from the injection section 30 into the mold cavity 13. It is also preferable to provide mold temperature control means, including a temperature control circuit (not shown), in the fixed mold 11 and the movable mold 12 to adjust them to a predetermined temperature. Alternatively, the casting mold 10 may be provided with a vacuum suction means (not shown), and this vacuum suction means may be used to directly vacuum the inside of the mold cavity 13.

[0025] The mold clamping unit 20 comprises a fixed platen 21 that supports the fixed mold 11, a movable platen 22 that supports the movable mold 12, and a mold clamping platen 23 that supports the mold clamping drive unit 25. The mold clamping drive unit 25 is a hydraulic drive means such as a hydraulic cylinder, and the mold clamping drive unit 25 and the movable platen 22 are connected via a cylinder rod 26. The fixed platen 21 and the mold clamping platen 23 are connected by a plurality of tie bars 24 that pass through the movable platen 22. The mold clamping control unit 27 operates the mold clamping drive unit 25, causing the movable platen 22 to move in the mold opening and closing direction via the cylinder rod 26, using the tie bars 24 as guides. Here, the movement in the direction in which the movable platen 22 and the fixed platen 21 move closer together, or the movement in the direction in which the movable mold 12 and the fixed mold 11 move closer together, is defined as the mold clamping operation, and the movement in the direction in which they move further apart is defined as the mold opening operation.

[0026] In Figure 1, the clamping drive unit 25 is shown as a hydraulic drive means such as a hydraulic cylinder, but it is not limited to this. For example, it may be an electric drive means using a ball screw mechanism that converts the rotational motion of an electric motor into linear motion, or a hybrid drive means that combines a hydraulic drive means and an electric drive means. Furthermore, multiple clamping drive units 25 may be arranged, or the clamping drive unit 25 may be arranged on the tie bar 24.

[0027] The injection unit 30 comprises a cylindrical injection sleeve 31 arranged horizontally, a plunger tip 32 that moves in the front-rear direction within the injection sleeve 31, a pouring port 34 that supplies molten metal M such as aluminum alloy into the injection sleeve 31, and an injection drive unit 36 ​​that controls the front-rear movement of the plunger tip 32. The tip of the injection sleeve 31 (opposite the pouring port 34) passes through the fixed platen 21 and the fixed mold 11 and is connected to the mold gate 14. Here, the sliding of the plunger tip 32 is defined as moving forward F when approaching the mold gate 14, forward movement when moving forward F, backward R when moving away from the mold gate 14, and backward movement when moving backward R.

[0028] Here, the plunger rod 33, which is connected to the plunger tip 32, is connected via a connecting portion 35 to the drive rod 37 of the injection drive unit 36 ​​of a hydraulic drive means such as a hydraulic cylinder. While the plunger tip 32 is waiting behind the pouring port 34, molten metal M is supplied from the pouring port 34 into the injection sleeve 31 using a molten metal supply means (not shown). The injection control unit 38 operates the injection drive unit 36, and the forward and backward movement of the plunger tip 32 is controlled via the drive rod 37, the connecting portion 35, and the plunger rod 33. The forward movement of the plunger tip 32 presses the molten metal M supplied into the injection sleeve 31, and it is injected and filled into the mold cavity 13 via the mold gate 14.

[0029] Furthermore, the injection sleeve 31 and plunger tip 32 are provided with cooling means (not shown) including a channel through which a cooling medium such as cooling water flows. In addition, it is preferable to apply a lubricant to the sliding surfaces of the injection sleeve 31 and plunger tip 32 in order to prevent galling damage due to strong contact between the plunger tip 32 and the injection sleeve 31, stabilize the sliding state, and suppress the adhesion of molten metal residue. Alternatively, the vacuum suction of the mold cavity 13 by a vacuum suction means provided in the casting mold 10 and injection filling may be performed simultaneously. Alternatively, a vacuum suction means (not shown) may be provided in the injection sleeve 31 to vacuum the inside of the injection sleeve 31 and indirectly vacuum the inside of the mold cavity 13 through the mold gate 14, or a combination of direct and indirect vacuum suction may be performed.

[0030] In Figure 1, the injection drive unit 36 ​​is shown as a hydraulic drive means such as a hydraulic cylinder, but it is not limited to this. For example, it may be an electric drive means using a ball screw mechanism that converts the rotational motion of an electric motor into linear motion, or a hybrid drive means that combines a hydraulic drive means and an electric drive means. Furthermore, in the case of a hydraulic drive means, it may also be equipped with a pressure accumulation means such as an accumulator.

[0031] Furthermore, although the injection sleeve 31 is positioned horizontally in Figure 1, it may be positioned arbitrarily within a range from horizontal to vertically downward relative to the mold cavity 13 and mold gate 14. Also, although molten metal M is supplied from the pouring port 34, it may be supplied by connecting the injection sleeve 31 to a melting furnace (not shown) that holds the molten metal M using a connecting means such as a molten metal supply pipe. Also, although the injection sleeve 31 is shown as a combination of a plunger tip 32, it may be a configuration in which pressurized gas is supplied to a sealed molten metal M holding furnace (not shown), and the molten metal M is injected and filled into the mold cavity 13 via a molten metal supply pipe, or a transport means such as a transport pump may be used to inject and fill the molten metal M instead of supplying pressurized gas.

[0032] The casting control unit 40 is connected to the mold clamping control unit 27 and the injection control unit 38. Based on preset casting conditions, it operates the mold clamping control unit 27 and the injection control unit 38 to control operations such as the clamping and opening / closing of the casting mold 10, and injection filling, which involves pressing molten metal M into the mold cavity 13, thereby performing casting. In addition, it sends operation commands to peripheral equipment such as molten metal supply means, cooling means, and vacuum suction means (not shown) to operate the casting apparatus 100 and manage the casting process.

[0033] The automatic determination unit 50 collects dummy casting waveform data transmitted from the casting apparatus 100 during dummy casting, and has the function of automatically determining whether to switch from dummy casting to production casting to manufacture planned production castings using the collected dummy casting waveform data. Further details will be explained using Figure 2. Note that in Figure 1, the casting control unit 40 and the automatic determination unit 50 are arranged separately, but for example, the casting control unit 40 and the automatic determination unit 50 may be integrated into a single structure.

[0034] Here, "sacrificial casting" is performed before starting production casting for the purpose of stabilizing the quality of production castings and adjusting casting conditions. Based on pre-set sacrifice casting conditions, injection filling of a predetermined amount of molten metal M into the mold cavity 13 and removal of the sacrifice casting that has cooled and solidified in the mold cavity 13 are repeated. This utilizes the heat of the molten metal M to preheat the casting mold 10, the injection sleeve 31 and plunger tip 32 of the injection unit 30 to a temperature suitable for casting. If the injection drive unit 36 ​​is a hydraulic drive means such as a hydraulic cylinder, the temperature of the hydraulic fluid of the hydraulic drive means is also preheated to a suitable state. Furthermore, casting conditions such as checking the flow state of the molten metal M in the mold cavity 13 and adjusting the casting pressure to a suitable level for adjusting the density of the molten metal M filled in the mold cavity 13 are checked. From the viewpoint of production efficiency, it is preferable to minimize sacrifice casting, and a highly accurate switching judgment from sacrifice casting to production casting is desired.

[0035] (Automatic determination unit 50) Next, the automatic determination unit 50 of the discard-type automatic determination system 200 shown in Figure 1 will be explained using Figure 2. First, as shown in Figure 2(a), the automatic determination unit 50 includes an AI creation mode 50S and an AI determination mode 50H. The AI ​​creation mode 50S includes a data transmission / reception unit 51, a data setting unit 52, and a learning unit 53. The AI ​​determination mode 50H includes a determination unit 54 and a display unit 55.

[0036] The data transmission / reception unit 51 in AI creation mode 50S is connected to the casting control unit 40 of the casting apparatus 100 and receives discard casting waveform data transmitted from the injection control unit 38, the mold clamping control unit 27, and the peripheral equipment 99 via the casting control unit 40. The peripheral equipment 99 includes a hot water supply means, a cooling means, and a vacuum suction means. The result of the switching determination is transmitted from the data transmission / reception unit 51 to the injection control unit 38, the mold clamping control unit 27, and the peripheral equipment 99 via the casting control unit 40.

[0037] Here, the test casting waveform data transmitted from the mold clamping control unit 27 includes, for example, the mold clamping force waveform, the drive torque waveform of the mold clamping drive unit 25, the vibration waveform of the mold clamping unit 20, the release force waveform of the mold clamping drive unit 25 at the initial stage of mold opening, the push force waveform of an extrusion means (not shown) built into the mold clamping unit 20, the mold opening and closing speed waveform, and the stress waveform generated in the multiple tie bars 24, with the time axis representing the period from the start to the end of one shot of test casting.

[0038] Furthermore, the discard casting waveform data transmitted from the injection control unit 38 includes, on a time axis from the start to the end of one shot of discard casting, for example, the forward speed waveform of the plunger tip 32 (referred to as the injection speed waveform), the pressing force waveform of the molten metal M by the plunger tip 32 (referred to as the casting pressure waveform), the time waveform required for injection filling to fill the mold cavity 13 with molten metal M (referred to as the injection time waveform), the position waveform of the plunger tip 32 (referred to as the injection position waveform), the drive torque waveform of the injection drive unit 36, the temperature waveform of the injection sleeve 31 or plunger tip 32, the temperature waveform of the molten metal M, the vibration waveform or vibration acceleration waveform during the forward and backward movement of the plunger tip 32, the injection speed switching time waveform (referred to as the acceleration waveform), the operation command waveform of the injection control unit 38 and the execution waveform of the injection drive unit 36, etc.

[0039] Furthermore, the discard casting waveform data transmitted from the peripheral equipment 99 includes, on a time axis from the start to the end of one discard casting shot, for example, the temperature waveform of the casting mold 10, the temperature waveform and flow rate waveform of the cooling medium used for the cooling means of the casting mold 10 and the injection unit 30, the vacuum waveform inside the mold cavity 13 or injection sleeve 31, the thermal expansion waveform of the casting mold 10, the waveform of the amount of release agent applied to the mold cavity 13 and the mold gate 14, the waveform of the deformation amount of the casting mold 10 during the clamping operation, the waveform of the molten metal transport time and transport amount waveform of the molten metal supply means, and environmental conditions such as temperature and humidity or atmospheric pressure.

[0040] As shown in Figure 2(b), the data setting unit 52 of the AI ​​creation mode 50S includes a waveform acquisition start switch 521, a measurement waveform selection switch 522, an axis selection switch 523, a learning interval setting switch 524, a learning model creation switch 525, and a learning model save switch 526. Note that the labels for the switches are omitted in Figure 2(b). Each switch also has a display panel EP that displays the setting status.

[0041] The waveform acquisition start switch 521 selects whether to start or stop the acquisition of discarded waveform data received by the data transmission / reception unit 51.

[0042] The measurement waveform selection switch 522 selects and sets waveform data from the test casting waveform data received by the data transmission / reception unit 51 to be used for determining the switch from test casting to production casting. The operation of the measurement waveform selection switch 522 is performed by the casting operator performing the test casting. For example, the casting operator may select and set the waveform data from the test casting waveform data that is deemed suitable for the switch determination, which is stored in the data setting unit 52 in past test castings. Hereinafter, assuming that the measurement waveform selection switch 522 has selected and set waveform data (called casting pressure waveform data) that shows the pressing force (called casting pressure) of the plunger tip 32 of the injection unit 30 onto the molten metal M during forward movement, the explanation will be based on this assumption.

[0043] The axis selection switch 523 selects and sets the axis on which to display the waveform data (for example, casting pressure waveform data) set by the measurement waveform selection switch 522. In this case, it is preferable to set the elapsed time from the start to the end of the forward movement of the plunger tip 32 during the injection filling of molten metal M into the mold cavity 13. Alternatively, it is preferable to set the forward position (referred to as the injection position) of the plunger tip 32 during injection filling. Note that the axis selection switch 523 may be set automatically based on the setting of the measurement waveform selection switch 522.

[0044] The learning interval setting switch 524 sets the learning interval of the waveform data (e.g., casting pressure waveform data) set by the measurement waveform selection switch 522 and saves it in the learning unit 53. Here, the data capacity of the discard casting waveform data collected in a discard casting is enormous, requiring an automatic determination unit 50 with an extremely large memory capacity, which would increase the size and cost of the discard casting automatic determination system 200. Furthermore, analyzing the enormous amount of data takes time, making instantaneous switching decisions difficult and resulting in low efficiency. Therefore, by setting only the necessary parts from the discard casting waveform data as the learning interval, the data capacity can be compressed, making the discard casting automatic determination system 200 smaller, reducing costs, and shortening the switching decision time. This learning interval setting may be done by a casting operator, for example, by referring to past discard castings. Alternatively, the analysis results may be set using analysis means such as flow analysis. By repeatedly performing switching decisions using the discard casting automatic determination system 200, the learning interval will be adjusted to a suitable range.

[0045] The learning model creation switch 525 defines the characteristics of the waveform data (for example, casting pressure waveform data) set by the measurement waveform selection switch 522, based on the learning interval set by the learning interval setting switch 524, in the learning unit 53 of the AI ​​creation mode 50S (this is called AI learning). The result of this AI learning is called an AI learning model, and the learning model save switch 526 assigns an identification number to the AI ​​learning model and saves it in the learning unit 53.

[0046] The judgment unit 54 of the AI ​​judgment mode 50H includes, as shown in Figure 2(c), a judgment mode selection switch 541, a teacher data specification switch 542, a threshold setting switch 543, a switching judgment start switch 544, and a judgment result save switch 545. Note that the labels for the switches are omitted in Figure 2(c). It also includes a display panel EP that shows the selection status of each switch.

[0047] The judgment mode selection switch 541 selects and sets whether to use a supervised learning program using a regression method or a classification method when making a decision on switching from test casting to production casting. For details, please refer to Figures 3 and 4.

[0048] The training data specification switch 542 is used to specify and set training data when making a switchover determination from sacrificial casting to production casting using a supervised learning program. Preferably, this training data uses an AI learning model created in AI creation mode 50S and stored in the learning unit 53. Alternatively, for example, the training data may be the waveform data of the sacrificial casting at the timing of the switchover from sacrificial casting to production casting, based on past casting performance. This timing may be, for example, the timing when the set number of sacrificial castings has been reached using a conventional switching determination means, or the timing when a switching determination is made by visual inspection by a casting engineer. Alternatively, it may be the timing determined by temperature measurement of the casting mold 10 or injection sleeve 31, etc. In any case, by repeatedly making switching determinations using the automatic sacrificial casting determination system 200, the training data is adjusted to a suitable state.

[0049] The threshold setting switch 543 sets a suitable numerical value as the threshold in accordance with the supervised learning program of the regression method or classification method set by the judgment mode selection switch 541. For example, the threshold may be calculated from past records of switching from sacrificial casting to production casting and set by the casting engineer performing the sacrificial casting process. Alternatively, the threshold may be set using analysis results obtained from analysis means such as flow analysis. Furthermore, the threshold may be adjusted based on the repeated switching judgment records of the judgment unit 54. Details will be explained using Figures 3 and 4.

[0050] The switching determination start switch 544 selects whether to start or stop the switching determination from dummy casting to production casting. When start is selected, the determination unit 54 uses a supervised learning program based on the setting of the determination mode selection switch 541 to compare the teacher data specified by the teacher data specification switch 542 with the dummy casting waveform data received by the data transmission / reception unit 51 to perform a switching determination.

[0051] The judgment result saving switch 545 assigns an identification number to the discard waveform data (referred to as judgment waveform data) at the timing of the switching judgment and saves it to the judgment unit 54. This judgment waveform data may be reused, for example, in the AI ​​learning model of AI creation mode 50S. This reuse is expected to further improve the accuracy of the AI ​​learning model. Alternatively, this judgment waveform data may be set as training data for AI judgment mode 50H using the training data specification switch 542. In this way, by repeatedly reusing the judgment waveform data in the AI ​​learning model of AI creation mode 50S and using it as training data for AI judgment mode 50H, it is expected that highly accurate switching judgments from discard casting to production casting can be made.

[0052] The display unit 55 displays the progress and result of the switching judgment performed by the judgment unit 54.

[0053] Herein, the present invention is characterized by using test-cast waveform data in AI creation mode 50S and AI judgment mode 50H. For example, the injection sleeve 31, etc., receives heat from the molten metal M and undergoes irregular thermal deformation, causing it to bend significantly (known as banana deformation), which increases the sliding resistance between the injection sleeve 31 and the plunger tip 32, disrupting the behavior of injection filling of the molten metal M into the mold cavity 13 and potentially causing significant fluctuations in casting quality. Alternatively, the behavior of injection filling can also be disrupted by galling, where the molten metal M penetrates the sliding surfaces of the injection sleeve 31 and the plunger tip 32. Furthermore, many factors are involved in test-casting, such as fluctuations in the amount of molten metal M supplied to the injection sleeve 31, the influence of molten metal residues such as oxides and impurities contained in the molten metal M, fluctuations in the composition of the molten metal M, aging of the driving means of the plunger tip 32, mold residue defects where a part of the casting remains in the mold cavity 13, and the state of release agent application to the mold cavity 13. Since the waveform data from the initial casting process includes all of these factors, using the initial casting waveform data for switching decisions is the most suitable approach.

[0054] In this way, the automatic test casting determination system 200, which connects the automatic determination unit 50 to the casting apparatus 100, is used to determine when to switch from test casting to production casting based on the test casting waveform data during test casting. The test casting waveform data can be used to check many factors that are said to affect the casting quality of the casting product all at once. This makes it possible to provide an automatic test casting determination system 200 that can automatically determine when to switch from test casting to production casting with high accuracy and efficiency. Furthermore, by creating an AI learning model that serves as an indicator for the switching determination in AI creation mode 50S, and using this AI learning model as training data with a supervised learning program to perform the switching determination in AI determination mode 50H, the accuracy of the switching determination is further improved, and it is expected that the production efficiency of casting will be increased.

[0055] (Automatic method for determining whether to discard a shot) Next, the automatic discard detection method using the discard detection system 200 shown in Figure 1 will be explained with reference to Figures 3 and 4. Figure 3 shows the control flow diagram of the automatic discard detection method, with Figure 3(a) showing the case when the supervised learning program for the regression method is selected using the detection mode selection switch 541 of the detection unit 54, and Figure 3(b) showing the case when the supervised learning program for the classification method is selected. Figure 4 shows an example of the display of the display unit 55, showing the progress and result of the switching detection performed by the detection unit 54, with Figures 4(a) to (c) showing the case when the supervised learning program for the regression method is selected, and Figure 4(d) showing the case when the supervised learning program for the classification method is selected.

[0056] First, as shown in Figure 3(a), the automatic determination method for discard casting when a supervised learning program of the regression method is selected using the determination mode selection switch 541 of the determination unit 54 will be explained. A discard casting process is performed based on the discard casting conditions set in advance by the casting control unit 40. Upon receiving a command from the casting control unit 40, the clamping control unit 27 of the clamping unit 20 operates the clamping drive unit 25 to clamp the casting mold 10 and form the mold cavity 13 and mold gate 14 (clamping process). It is preferable to apply a release agent or the like to the mold cavity 13 before the clamping process. Alternatively, a vacuum suction means (not shown) may be used to vacuum-suction the inside of the mold cavity 13. Next, a predetermined amount of molten metal M is supplied from the pouring port 34 of the injection unit 30 into the injection sleeve 31 using a molten metal supply means (not shown) (molten metal supply process). Upon receiving a command from the casting control unit 40, the injection control unit 38 operates the injection drive unit 36 ​​to move the plunger tip 32 forward, injecting and filling the mold cavity 13 with molten metal M from the injection sleeve 31 via the mold gate 14 (sacrificial injection process). Subsequently, the cooled and solidified sacrificial casting is removed from the mold cavity 13 (removal process).

[0057] In this test casting process, the automatic determination unit 50 makes a decision to switch from the test casting process to the production casting process (automatic determination process) based on test casting waveform data transmitted from the mold clamping control unit 27, injection control unit 38 and peripheral equipment 99 via the casting control unit 40, and the process transitions from test casting to production casting. The test casting process is a preparatory process performed before starting production casting for the purpose of stabilizing the quality of production castings and adjusting casting conditions, and it involves repeating the test injection process and the removal process. The production casting process refers to the process of injecting and filling molten metal M into the mold cavity 13 to cast production castings based on production conditions set in advance by the casting control unit 40.

[0058] Furthermore, the automatic judgment process comprises an AI creation process and an AI judgment process. The AI ​​creation process comprises a waveform data collection process, a data setting process, and a learning model creation process. The AI ​​judgment process comprises a training data setting process, a switching judgment process, and a display process.

[0059] First, the waveform data acquisition process starts the waveform acquisition start switch 521 of the data setting unit 52, and the data transmission / reception unit 51 acquires the test shot waveform data transmitted from the mold clamping control unit 27, the injection control unit 38 and peripheral equipment 99 via the casting control unit 40. The acquired test shot waveform data is transferred to the data setting unit 52 and stored.

[0060] Following the waveform data acquisition process, the data setting process involves selecting waveform data from the test casting waveform data collected in the waveform data acquisition process to be used for determining the switch from test casting to production casting, using the measurement waveform selection switch 522 of the data setting unit 52. While the selection of waveform data is assumed to be performed by the casting operator performing the test casting process, it may also be automatically selected based on past test casting process performance. Here, for example, let's assume that the measurement waveform selection switch 522 has selected and set waveform data (referred to as casting pressure waveform data) indicating the pressing force (referred to as casting pressure) of the plunger tip 32 of the injection unit 30 onto the molten metal M during forward movement, and the following explanation will be based on this assumption.

[0061] Furthermore, the data setting process involves selecting and setting the axis of the selected waveform data (for example, casting pressure waveform data) using the axis selection switch 523 of the data setting unit 52. In this case, it is preferable to set the elapsed time from the start to the end of the forward movement of the plunger tip 32 during the injection filling of molten metal M into the mold cavity 13. Alternatively, it is preferable to set the forward position of the plunger tip 32 during injection filling (referred to as the injection position). Note that the axis selection switch 523 may be set automatically based on the setting of the measurement waveform selection switch 522. Here, the selected waveform data (for example, casting pressure waveform data) is shown in the display unit 55 as casting pressure waveform data HK with the horizontal axis as the injection position and the vertical axis as the casting pressure, as shown in Figure 4(a).

[0062] Following the data setting process, the learning model creation process sets the learning interval for the selected waveform data (e.g., casting pressure waveform data HK) using the learning interval setting switch 524 of the data setting unit 52. The set learning interval is stored in the learning unit 53. Here, the data volume of the discard casting waveform data collected in a discard casting, including the selected waveform data (e.g., casting pressure waveform data HK), is enormous, requiring an automatic judgment unit 50 with an extremely large memory capacity, resulting in a larger discard casting automatic judgment system 200 and increased costs. Furthermore, analyzing the enormous amount of data takes time, making instantaneous switching judgment difficult and resulting in low efficiency. Therefore, by setting only the necessary parts from the discard casting waveform data as the learning interval, the data volume can be compressed, making the discard casting automatic judgment system 200 smaller, reducing costs, and shortening the switching judgment time.

[0063] The learning interval is set, for example, based on past performance of the sacrificial casting process, as the range in which changes occur in the sacrificial casting waveform data, and the learning interval (G1, G2) is set at the injection position as shown in Figure 4(a). Alternatively, the results of analysis using analysis methods such as flow analysis may be set as the learning interval (G1, G2). By repeatedly performing switching judgments using the sacrificial casting automatic judgment system 200, the learning interval (G1, G2) is adjusted to a suitable range.

[0064] Furthermore, the learning model creation process involves activating the learning model creation switch 525 of the data setting unit 52, and having the learning unit 53 define the features of the selected waveform data (for example, casting pressure waveform data HK) within the set learning intervals (G1, G2) (this is called AI learning). The result of this AI learning is called the AI ​​learning model, and the learning model save switch 526 of the data setting unit 52 assigns an identification number to the AI ​​learning model and saves it in the learning unit 53. This AI learning, for example, uses a deep learning image analysis method to define the features of the casting pressure waveform data HK. As a result, as shown in Figure 4(a), the features are defined such that in a test casting process limited to casting pressure P1, the casting pressure increases within the range of the learning intervals (G1, G2), the test injection process ends, and the injection position in learning interval G1 is switched.

[0065] Following the learning model creation process, the training data setting process uses a supervised learning program of the regression method to set training data for determining the switch from sacrificial casting to production casting, by specifying the training data specification switch 542 of the determination unit 54. Preferably, this training data uses the AI ​​learning model created in AI creation mode 50S and stored in the learning unit 53. Alternatively, for example, the sacrificial casting waveform data at the timing of the switch from sacrificial casting to production casting, based on past sacrificial casting process results, may be used as training data. This timing may be, for example, the timing when the set number of sacrificial castings is reached using a conventional switching determination means, or the timing when a switching determination is made by visual inspection by a casting engineer. Alternatively, it may be the timing determined by temperature measurement of the casting mold 10 or injection sleeve 31, etc. In any case, by repeatedly performing switching determinations using the automatic sacrificial casting determination system 200, the training data is adjusted to a suitable state. Here, we will continue the explanation assuming that the casting pressure waveform data HK and learning intervals (G1, G2) that serve as the basis for the AI ​​learning model are specified as training data.

[0066] Following the teacher data setting process, the switching determination start switch 544 of the determination unit 54 is activated, and the switching determination process is performed by the determination unit 54. Here, the switching determination process when the supervised learning program of the regression method is selected will be described using FIG. 4(b). First, in the learning intervals (G1, G2), the casting pressure waveform data HK specified in the teacher data is compared with the casting pressure waveform data (H1 to H3) in the discard shot waveform data collected in the waveform data collection process and set in the data setting process, and the difference amount is quantified (referred to as a determination score value) based on the rules preset in the determination unit 54.

[0067] For example, since the casting pressure waveform data H1 at the start of the discard casting process is in a low-temperature state such as the casting mold 10, the molten metal M is rapidly cooled, and the discard injection process ends at the injection position S1 at a short distance, showing the determination score value K1(G1 - S1). By repeating the discard casting process, the casting mold 10 etc. are heated by the molten metal M and the temperature rises, and the cooling of the molten metal M during the discard injection process is delayed. As a result, the injection positions (S2, S3) indicating the end of the discard injection process of the casting pressure waveform data (H2, H3) gradually become longer in distance (S1 < S2 < S3), and accordingly, the determination score values (K2, K3) gradually become smaller (K1 > K2 > K3). In FIG. 4(b), for simplicity of explanation, the number of times of the discard casting process is set to 3 times (casting pressure waveform data H1 to H3), but in reality, it is more.

[0068] Furthermore, as shown in FIG. 4(c), in the switching determination process, the determination unit 54 counts the determination score values (K1 to K3) for each discard casting shot number of the discard casting process. With the horizontal axis being the discard casting shot number and the vertical axis being the determination score value, it is displayed on the display unit 55. Note that FIG. 4(c) counts a plurality of determination score values according to the actual number of times of the discard casting process and displays them as black dots. Furthermore, in the switching determination process, the determination unit 54 performs regression analysis on the plurality of counted determination score values and overwrites and displays the regression analysis value KB.

[0069] Here, the selected casting pressure waveform data and other test casting waveform data include many factors such as changes in the sliding resistance of the injection sleeve 31 and plunger tip 32 due to thermal deformation and galling, fluctuations in the supply amount, temperature, and composition of molten metal M, the influence of molten metal residue, aging changes in the injection unit 30, and fluctuations in mold residue defects and lubrication application conditions in the mold cavity 13. As a result, the test casting waveform data may fluctuate suddenly and significantly, and consequently, the judgment score value also fluctuates significantly. When the judgment score value fluctuates significantly, the accuracy of the switching judgment using the judgment score value decreases, which is undesirable. In contrast, the supervised learning program of the regression method performs regression analysis on the judgment score value to obtain the regression analysis value KB, and by using this regression analysis value KB, sudden fluctuations can be avoided and the accuracy of the switching judgment can be improved. Furthermore, the timing of the switching judgment can be predicted from the regression analysis value KB, and preparation for the production casting process can be started in parallel with the test casting process, thereby improving the efficiency of the casting work.

[0070] Furthermore, the switching determination process compares the regression analysis value KB with the threshold R1 set by the threshold setting switch 543 of the determination unit 54, and the determination unit 54 makes a decision on switching from sacrificial casting to production casting. For example, as shown in Figure 4(c), the sacrificial casting shot A1 at the timing when the regression analysis value KB ≤ threshold R1 is set as the switching determination. Assuming that the sacrificial casting shot A2 is the switching determination of the conventional switching determination means, this shows that the sacrificial casting shot A3 (A2-A1) can be shortened in the sacrificial casting automatic determination method of the present invention. Figures 4(a) to 4(c) are displayed on the display unit 55 as a display process. When the determination unit 54 confirms that the sacrificial casting shot A1 has been determined to be a switching determination, the result of the switching determination is transmitted from the data transmission / reception unit 51 to the casting control unit 40, then to the mold clamping control unit 27, the injection control unit 38, and peripheral equipment 99, and the casting apparatus 100 finishes the sacrificial casting process and starts the production casting process.

[0071] Here, the threshold R1 may be calculated, for example, from past switching judgment results, or it may be set using analysis results obtained from analytical methods such as flow analysis. Furthermore, the threshold R1 is adjustable. For example, if the casting quality of the castings in the production casting process is stable after switching from the sacrificial casting process to the production casting process, the threshold R1 is adjusted in the direction of increasing. As a result, the number of casting shots in the sacrificial casting process can be reduced, and the efficiency of casting can be increased. On the other hand, if the casting quality is not stable, the threshold R1 is adjusted in the direction of decreasing. As a result, the number of casting shots in the sacrificial casting process will increase, but the casting quality in the production casting process can be stabilized.

[0072] In this automatic discard casting determination method using an automatic discard casting determination system 200 in which an automatic determination unit 50 is connected to a casting apparatus 100, a determination to switch from discard casting to production casting is made based on the discard casting waveform data during the discard casting. The discard casting waveform data can be used to check many factors that are said to affect the casting quality of the casting product all at once. This makes it possible to provide an automatic discard casting determination method that can automatically determine the switch from discard casting to production casting with high accuracy and efficiency. Furthermore, since an AI learning model that serves as an indicator for the switch determination is created in the AI ​​creation process, and this AI learning model is used as training data to perform the switch determination in the AI ​​determination process using a supervised learning program, the accuracy of the switch determination can be further improved.

[0073] Furthermore, by using a supervised learning program for regression methods, even if there are sudden fluctuations in the test casting waveform data, the fluctuations can be appropriately stabilized, further improving the accuracy of the switching decision. At the same time, the timing of the switching decision can be appropriately predicted, allowing preparation for the production casting process to proceed simultaneously with the test casting process. This makes it possible to increase the production efficiency of casting.

[0074] Next, as shown in Figure 3(b), we will explain the automatic discard casting method when a supervised learning program for the classification method is selected using the judgment mode selection switch 541 of the judgment unit 54. Note that the discard casting process, the waveform data collection process, data setting process and learning model creation process of the AI ​​creation process, and the teacher data setting process and display process of the AI ​​judgment process are substantially the same as in Figure 3(a), so we will omit their explanation and focus on explaining the switching judgment process of the AI ​​judgment process, which differs from that in Figure 3(a).

[0075] First, as shown in Figure 4(d), the data setting process involves selecting the waveform data (casting pressure waveform data HB) used for determining the switch from sacrificial casting to production casting using the measurement waveform selection switch 522 of the data setting unit 52. Similarly, the learning intervals (G3, G4) of the casting pressure waveform data HB are set using the learning interval setting switch 524 of the data setting unit 52. These learning intervals (G3, G4) are adjusted to a suitable range by repeatedly performing the switch determination using the sacrificial casting automatic determination system 200. For example, the fluctuation range of past switching determinations from the sacrificial casting process to the production casting process may be set as the learning intervals (G3, G4).

[0076] Furthermore, in the training data setting process, in the automatic throwaway judgment method using a supervised learning program for the classification method, the learning interval (G3, G4) and the casting pressure waveform data HB are set as training data using the training data specification switch 542 of the judgment unit 54. Similarly, the learning interval (G3, G4) is set as the threshold R2 using the threshold setting switch 543 of the judgment unit 54.

[0077] Next, in the switching determination process when a supervised learning program for the classification method is selected, the casting pressure waveform data HB specified as training data is compared with the casting pressure waveform data (H4~H6) from the discard casting waveform data collected in the waveform data acquisition process and set in the data setting process, at the learning interval (G3, G4) or threshold R2, and a switching determination is made from the discard casting process to the production casting process.

[0078] For example, in the initial stage of the start of the thump casting process, the casting pressure waveform data H4 shows that since the casting mold 10 and the like are in a low-temperature state, the molten metal M is rapidly cooled, and the thump injection process ends at the injection position S4 with a short distance, and the learning sections (G3, G4) or the threshold value R2 have not been reached. By repeating the thump casting process, the casting mold 10 and the like are heated by the molten metal M and the temperature rises, and the cooling of the molten metal M during the thump injection process is delayed. As a result, the injection positions (S5, S6) indicating the end of the thump injection process of the casting pressure waveform data (H5, H6) gradually become longer (S4 < S5 < S6). Note that the injection position S5 of the casting pressure waveform data H5 has not reached the learning sections (G3, G4) or the threshold value R2 either. At the injection position S6 of the casting pressure waveform data H6, it is shown that the learning sections (G3, G4) or the threshold value R2 have been reached for the first time, and the thump casting shot of this casting pressure waveform data H6 is used as a switching determination from the thump casting process to the production casting process, and the thump casting process is ended and the production casting process is started.

[0079] When the switching determination process selects the supervised learning program of the classification method, it is an effective means for switching determination that is effective when there are no sudden fluctuations and it is relatively stable in the waveform data selected by the measurement waveform selection switch 522 of the data setting unit 52 for switching determination from thump casting to production casting. In addition, since it does not require counting of the determination score values and calculation of regression analysis values performed in the switching determination process using the supervised learning program of the regression method, it enables miniaturization and cost reduction of the thump automatic determination system 200. Furthermore, the thump waveform data can be collectively confirmed for many factors that are considered to affect the casting quality of the cast product. Thereby, it is possible to provide a thump automatic determination method capable of automatically and efficiently performing the switching determination from thump casting to production casting with high accuracy. When the selected waveform data has low stability or when it is desired to predict the timing of the switching determination, it is preferable to select the switching determination process using the supervised learning program of the regression method.

[0080] Although preferred embodiments of the present invention have been described above, the technical scope of the present invention is not limited to the embodiments described above. Various modifications or improvements can be made to the above embodiments.

[0081] For example, the switching determination process using the supervised learning program of the classification method shown in Figure 4(d) has a threshold R2 (or learning intervals G3, G4) set, but it is not limited to this, and the switching determination may be made without setting a threshold R2 (or learning intervals G3, G4). In this case, it is preferable that the switching determination is made when the injection position (S4~S6) indicating the end of the test injection process reaches or exceeds the learning interval G3.

[0082] Furthermore, for example, in the switching determination process using the supervised learning program of the regression method shown in Figure 4(b), a pressure limit is set on the injection drive unit 36 ​​of the injection unit 30 to achieve a casting pressure P1, and a test casting process is performed. However, as shown in Figure 5(a), the test casting process may also be performed with the pressure limit set on the injection drive unit 36 ​​removed. In Figure 5(a), the horizontal axis represents the injection position (or injection elapsed time), and the vertical axis represents the metal pressure measured by the injection drive unit 36. Within the learning interval (G5, G6), the metal pressure (MP1 to MP4) of the metal pressure waveform data (M1 to M4) shows the maximum value. Here, for the same reasons as in Figure 4(b), the metal pressure gradually decreases from the start of the test casting process towards the switching determination (MP1>MP2>MP3>MP4), so it may also be expanded into an AI learning model and a switching determination may be performed using the supervised learning program of the regression method.

[0083] Alternatively, as shown in Figure 5(b), the metal pressure (MP5 to MP8) of the metal pressure waveform data (M5 to M8) as it passes through the learning interval G7 may be measured (called fixed-point measurement), and the switching determination process may be performed using a supervised learning program of a classification method or regression method. The advantage of this fixed-point measurement is that the capacity of the discard waveform data collected and stored by the automatic determination unit 50 can be reduced, enabling miniaturization of the automatic determination unit 50, cost reduction, and a reduction in the time required for switching determination. [Explanation of Symbols]

[0084] 100 Casting apparatus 200 Automatic Discard Shot Detection System 10 Casting molds 11 Fixed mold 12. Movable molds 13 Mold Cavity 14 Mold Gate 20 Mold clamping part 21 Fixed plate 22 Movable plate 23 Mold clamping board 24 Tie Bar 25-type clamping drive unit 26 Cylinder rod 27 Type clamping control unit 30 Injection part 31 Injection Sleeve 32 plunger tips 33 Plunger Rod 34 pouring spouts 35 Connecting part 36 Injection drive unit 37 Drive Rod 38 Injection control unit 40 Casting Control Unit 50 Automatic judgment section 50S AI creation mode 50H AI Judgment Mode 51 Data transmission / reception unit 52 Data Setting Section 521 Waveform acquisition start switch 522 Measurement waveform selection switch 523 Axis selection switch 524 Learning interval setting switch 525 Learning Model Creation Switch 526 Learning Model Save Switch EP display panel 53 Learning Department 54 Judgment section 541 Judgment Mode Selection Switch 542 Training data specification switch 543 Threshold setting switch 544 Switching judgment start switch 545 Switch to save judgment result 55 Display section 99 Peripheral facilities HK, HB waveform data (casting pressure waveform data, training data) H1~H6 Casting Pressure Waveform Data G1-G7 Learning Interval S1~S6 injection position K1~K3 Judgment Score Value P1 Casting pressure KB regression analysis values R1, R2 thresholds A1-A3 Discarded Casting Shots M1~M8 Metal Pressure Waveform Data MP1~MP8 Metal Pressure M molten metal F forward R rear

Claims

1. In an automated sacrificial casting determination system, which determines whether to switch from sacrificial casting, in which molten metal is repeatedly injected and filled into the mold cavity based on pre-set sacrificial casting conditions, to production casting, in which molten metal is injected and filled into the mold cavity based on pre-set production conditions to cast production products, The system includes an automatic determination unit that automatically performs the switching determination using the waveform data of the aforementioned test casting, The automatic determination unit is A data transmission / reception unit that receives the discard waveform data and transmits the result of the switching determination, A data setting unit that selects waveform data to be used for the switching determination from the discarded waveform data and sets the axis of said waveform data, A learning unit sets a learning interval for the waveform data, defines features through AI learning within the set learning interval, and creates and saves waveform data that serves as the basis for comparison in the switching determination. An automatic discard-shot determination system characterized by comprising a determination unit that performs the switching determination by comparing the aforementioned reference waveform data with the discard-shot waveform data.

2. The automatic discard judgment system according to claim 1, further comprising a display unit for displaying the data used for the switching judgment.

3. An automatic method for determining discarded shots using an automatic discarded shot determination system, In this automated test casting determination method, a determination is made to switch from a test casting process, in which molten metal is repeatedly injected and filled into the mold cavity based on pre-set test casting conditions, to a production casting process, in which molten metal is injected and filled into the mold cavity to cast production products based on pre-set production conditions. The system includes an automatic determination process that automatically performs the switching determination using the waveform data of the test casting process. The aforementioned automatic determination process is: A waveform data acquisition step which includes receiving the discard waveform data and transmitting the result of the switching determination, A data setting step involves selecting waveform data to be used for the switching determination from the discarded waveform data and setting the axis of said waveform data. A learning process that involves setting a learning interval for the waveform data, defining features through AI learning within the set learning interval, and creating and saving waveform data that serves as the basis for comparison in the switching judgment. An automatic discard detection method characterized by comprising a switching detection step of comparing the aforementioned reference waveform data with the discarded waveform data and performing the switching detection.

4. The automatic discard judgment method according to claim 3, further comprising a display step of displaying the data used for the switching judgment.

5. The automatic discard detection method according to claim 3 or 4, wherein the switching detection step counts the difference between the reference waveform data and the discarded waveform data as a detection score value, and performs the switching detection by comparing the regression analysis value obtained by regression analysis of the detection score value with a preset threshold value.

6. The automatic discard detection method according to claim 3 or 4, wherein the switching determination step involves comparing the reference waveform data with the discard waveform data and evaluating whether the discard waveform data has approached the reference waveform data to perform the switching determination.

Images