Quality prediction device
The quality prediction device addresses inaccuracies in existing systems by using a calculation model to account for varying parameters, ensuring precise casting quality assessment.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- TOYOTA JIDOSHA KK
- Filing Date
- 2025-09-08
- Publication Date
- 2026-06-01
AI Technical Summary
Existing prediction devices for casting product quality fail to account for varying parameters during each casting process, leading to inaccurate predictions.
A quality prediction device that incorporates a storage device and execution device, utilizing a calculation model to predict casting defects based on input values such as mold temperatures, vacuum pressures, and cooling water flow rates, enabling accurate quality assessment.
The device can predict casting product quality by considering fluctuating parameters, providing reliable quality predictions despite variations in casting conditions.
Smart Images
Figure 2026089660000001_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a quality prediction device.
Background Art
[0002] Patent Document 1 describes a prediction device for predicting the quality of a casting product. The prediction device includes an execution device and a storage device. The storage device stores a learned model obtained by machine learning. When a plurality of input values are input, the learned model outputs the probability of occurrence of shrinkage cavities as an output value.
Prior Art Documents
Patent Documents
[0003]
Patent Document 1
Summary of the Invention
Problems to be Solved by the Invention
[0004] The prediction device as described in Patent Document 1 cannot reflect the influence of parameters that vary each time a casting product is cast. Therefore, the prediction device cannot predict the quality of a casting product according to the influence of parameters that vary when the casting product is cast.
Means for Solving the Problems
[0005] A quality prediction device for solving the above problems predicts the quality of a cast product cast by a die-casting apparatus comprising: a fixed mold; a movable mold that is tightened to the fixed mold and partitions the cavity together with the fixed mold; a sleeve having an internal space connected to the cavity; a plunger that injects molten metal poured into the internal space into the cavity; and a vacuum tank for reducing the pressure of the cavity. The quality prediction device comprises a storage device and an execution device, the storage device storing a calculation model, the calculation model being the temperature of the fixed mold at the start of injection, which is the temperature of the fixed mold when the plunger starts injection, and the temperature of the fixed mold when the plunger starts injection and the temperature of the fixed mold when the plunger starts injection and the temperature of the fixed mold when the plunger starts injection When a plurality of input values are input, including the temperature of the moving mold at the start of pouring, which is the temperature of the moving mold when pouring is performed, the tank vacuum pressure of the vacuum tank for reducing the pressure of the cavity, the mold clamping completion position which indicates the position where the clamping operation of the moving mold to the fixed mold is completed, the tip galling pressure which is the pressure applied to the plunger tip when the plunger tip of the plunger slides, and the cooling water flow rate which is the flow rate of cooling water for cooling the fixed mold and the moving mold, the execution device outputs a parameter as an output value that indicates the amount of casting defects that are predicted to occur in the casting product, and the execution device predicts the quality of the casting product based on the output value output by the calculation model. [Effects of the Invention]
[0006] The above-described quality prediction device can obtain output values corresponding to multiple input values, including the fixed mold temperature at the start of injection, the moving mold temperature at the start of pouring, the tank vacuum pressure, the mold clamping completion position, the chip galling pressure, and the cooling water flow rate. Therefore, even if the quality of the cast product changes with each casting according to these input values, the device can predict the quality of the cast product according to these input values. [Brief explanation of the drawing]
[0007] [Figure 1] Figure 1 is a schematic diagram showing a die-casting system according to the first embodiment. [Figure 2]Figure 2 is a flowchart showing the series of processes performed by the quality prediction device in Figure 1 to predict the quality of a casting product. [Figure 3] Figure 3 is a schematic diagram showing a die-casting system according to the second embodiment. [Figure 4] Figure 4 is a graph illustrating the ejection speed of the plunger controlled by the control device in Figure 3. [Figure 5] Figure 5 is a flowchart showing the series of processes performed by the quality prediction device in Figure 3 to predict the quality of a cast product. [Modes for carrying out the invention]
[0008] <First Embodiment> In the following, a die-casting system equipped with a quality prediction device according to the first embodiment will be described with reference to the drawings.
[0009] <Overview of the die-casting system> As shown in Figure 1, the die-casting system 10 includes a die-casting device 20 and a quality prediction device 90.
[0010] The die-casting apparatus 20 casts aluminum alloy products by die-casting. The die-casting apparatus 20 includes a mold 30, a sleeve 40, a plunger 50, a vacuum tank 60, and a control device 70.
[0011] The mold 30 comprises a fixed mold 31 and a movable mold 32. The movable mold 32 is clamped to the fixed mold 31 by a movable mechanism (not shown). With the movable mold 32 clamped to the fixed mold 31, the mold 30 defines the cavity C. That is, the movable mold 32, by being clamped to the fixed mold 31, defines the cavity C together with the fixed mold 31.
[0012] Cavity C has the shape of the cast product to be cast by the die-casting apparatus 20. When the movable mechanism completes the tightening operation of the movable mold 32 with a specified pressure, the movable mold 32 is positioned at the target position. If there is no molten metal in cavity C and the movable mold 32 is tightened by the movable mechanism, the position where the movable mold 32 is tightened to the fixed mold 31 becomes the target position.
[0013] Although not shown in the diagram, the inside of the mold 30 is divided into multiple cooling water channels for circulating cooling water. As the cooling water flows through the cooling water channels, heat exchange occurs between the cooling water and the mold 30. In this way, the mold 30 is cooled by the cooling water.
[0014] The sleeve 40 is connected to the mold 30. The sleeve 40 is cylindrical. Therefore, the sleeve 40 has an internal space. The internal space of the sleeve 40 is connected to the cavity C. The plunger 50 is located in the internal space of the sleeve 40.
[0015] The plunger 50 pushes the molten metal from the internal space of the sleeve 40 into the cavity C. The plunger 50 has a plunger body 51 and a plunger tip 52. The plunger body 51 is rod-shaped. The plunger tip 52 is connected to the tip of the plunger body 51.
[0016] The plunger tip 52 is slidable within the internal space of the sleeve 40 as the plunger body 51 moves in the direction in which the sleeve 40 extends. Therefore, when the plunger body 51 is moved in the direction in which the sleeve 40 extends by an injection actuator (not shown), the plunger tip 52 pushes the molten metal present in the internal space of the sleeve 40 into the cavity C.
[0017] The vacuum tank 60 reduces the pressure in cavity C. The vacuum tank 60 is connected to cavity C by a connecting pipe. Although not shown in the diagram, the vacuum tank 60 reduces the pressure in cavity C by controlling a pressure reducing valve provided in the connecting pipe.
[0018] The control device 70 controls the die-casting device 20 to perform casting. After a mold release agent is applied inside the mold 30, the control device 70 controls the movable mechanism so as to clamp the movable mold 32 to the fixed mold 31.
[0019] When the clamping operation of the movable mold 32 to the fixed mold 31 is completed, the control device 70 controls the injection actuator so as to pour molten metal into the internal space of the sleeve 40. When the pouring of molten metal is completed, the control device 70 controls a pressure reducing valve (not shown) so as to reduce the pressure in the cavity C by the vacuum tank 60.
[0020] When the pressure reduction in the cavity C is completed, the control device 70 starts the injection of the plunger 50 so as to fill the cavity C with molten metal. When the filling of the cavity C with molten metal is completed, the control device 70 controls the movable mechanism so as to separate the movable mold 32 from the fixed mold 31 after a specified time has elapsed.
[0021] The die-casting system 10 includes a fixed mold temperature sensor 81, a movable mold temperature sensor 82, a tank pressure sensor 83, a position sensor 84, a chip pressure sensor 85, and a flow rate sensor 86.
[0022] The fixed mold temperature sensor 81 detects the fixed mold temperature P1. The fixed mold temperature P1 is the temperature of the fixed mold 31. The fixed mold temperature sensor 81 transmits the fixed mold temperature P1 at the start of injection, which is the fixed mold temperature P1 when the plunger 50 starts injection, to the quality prediction device 90.
[0023] The movable mold temperature sensor 82 detects the movable mold temperature P2. The movable mold temperature P2 is the temperature of the movable mold 32. The movable mold temperature sensor 82 transmits the movable mold temperature P2 at the start of pouring, which is the movable mold temperature P2 when molten metal is poured into the internal space of the sleeve 40, to the quality prediction device 90.
[0024] The tank pressure sensor 83 detects the tank vacuum pressure P3 of the vacuum tank 60. The tank pressure sensor 83 transmits the tank vacuum pressure P3 when the depressurization of cavity C is completed to the quality prediction device 90.
[0025] The position sensor 84 detects the mold clamping completion position P4, which indicates the position where the clamping operation of the moving mold 32 to the fixed mold 31 has been completed. The mold clamping completion position P4 is, for example, the distance from the target position to the position of the moving mold 32 when the clamping operation is completed. The position sensor 84 transmits the mold clamping completion position P4 to the quality prediction device 90.
[0026] The tip pressure sensor 85 detects the tip galling pressure P5. The tip galling pressure P5 is the pressure applied to the plunger tip 52 when the plunger tip 52 slides against the sleeve 40. The tip pressure sensor 85 transmits the tip galling pressure P5 to the quality prediction device 90.
[0027] The flow sensor 86 detects the cooling water flow rate P6. The cooling water flow rate P6 is the flow rate of the cooling water used to cool the mold 30 in the cooling water channel of the mold 30. For example, the cooling water flow rate P6 is the flow rate at the outlet along the path of the cooling water channel. The flow sensor 86 transmits the cooling water flow rate P6 to the quality prediction device 90.
[0028] The quality prediction device 90 obtains the fixed mold temperature P1 at the start of injection from the fixed mold temperature sensor 81. The quality prediction device 90 obtains the mobile mold temperature P2 at the start of pouring from the mobile mold temperature sensor 82. The quality prediction device 90 obtains the tank vacuum pressure P3 when the depressurization of cavity C is completed from the tank pressure sensor 83.
[0029] The quality prediction device 90 obtains the mold clamping completion position P4 from the position sensor 84. The quality prediction device 90 obtains the chip galling pressure P5 from the chip pressure sensor 85. The quality prediction device 90 obtains the cooling water flow rate P6 from the flow rate sensor 86.
[0030] The quality prediction device 90 comprises an execution device 91 and a storage device 92. The execution device 91 is a CPU. The storage device 92 is memory. The storage device 92 stores a prediction program PR for predicting the quality of the cast products cast by the die-casting device 20. By executing the prediction program PR on the execution device 91, the quality prediction device 90 predicts the quality of the cast products cast by the die-casting device 20.
[0031] The storage device 92 stores the calculation model MD used in conjunction with the execution of the prediction program PR. When the calculation model MD receives multiple input values, including multiple input values, it outputs the total volume TV of casting defects predicted to occur in the cast product cast by the die-casting device 20 as an output value. The total volume TV of casting defects is a parameter that indicates the amount of casting defects.
[0032] Multiple input values include the fixed mold temperature P1 at the start of injection, the moving mold temperature P2 at the start of pouring, the tank vacuum pressure P3, the mold clamping completion position P4, the tip galling pressure P5, and the cooling water flow rate P6.
[0033] The calculation model MD is predetermined through testing and simulation. The calculation model MD is a machine learning model determined by machine learning using training data. The training data used in machine learning is the total volume TV of casting defects in the cast product, obtained using measuring instruments.
[0034] When the casting of the cast product by the die-casting apparatus 20 is completed, the execution device 91 starts executing the prediction program PR. In other words, the execution device 91 starts executing the prediction program PR each time a cast is made.
[0035] As shown in Figure 2, when the execution device 91 starts executing the prediction program PR, the execution device 91 first performs the process in step S11. In step S11, the execution device 91 obtains the fixed mold temperature P1 at the start of injection, the moving mold temperature P2 at the start of pouring, the tank vacuum pressure P3, the mold clamping completion position P4, the tip galling pressure P5, and the cooling water flow rate P6. After that, the execution device 91 proceeds to step S12.
[0036] In step S12, the execution device 91 inputs several input values into the calculation model MD, including the fixed mold temperature P1 at the start of injection, the moving mold temperature P2 at the start of pouring, the tank vacuum pressure P3, the mold clamping completion position P4, the tip galling pressure P5, and the cooling water flow rate P6. After that, the execution device 91 proceeds to step S13.
[0037] In step S13, the execution device 91 obtains the total volume TV of the casting defects in the cast product as an output value output from the calculation model MD. The larger the value of the total volume TV of the casting defects, the greater the amount of casting defects. After that, the execution device 91 proceeds to step S14.
[0038] In step S14, the execution device 91 predicts the quality of the cast product based on the output values output from the acquired calculation model MD. For example, the execution device 91 calculates a quality parameter that increases as the total volume TV of casting defects increases. A smaller value for the quality parameter indicates better quality. After that, the execution device 91 completes this series of processes.
[0039] <Effects of the First Embodiment> (1) The quality prediction device 90 predicts the quality of the cast product to be cast by the die casting device 20. The quality prediction device 90 comprises a storage device 92 and an execution device 91. The storage device 92 stores a calculation model MD. When multiple input values are input, the calculation model MD outputs a total volume TV of casting defects as an output value, which is a parameter indicating the amount of casting defects that are predicted to occur in the cast product. The execution device 91 predicts the quality of the cast product based on the output value output by the calculation model MD. The multiple input values include the fixed mold temperature P1 at the start of injection, the moving mold temperature P2 at the start of pouring, the tank vacuum pressure P3, the mold clamping completion position P4, the chip galling pressure P5, and the cooling water flow rate P6.
[0040] The calculation model MD outputs output values corresponding to multiple input values, including the fixed mold temperature P1 at the start of injection, the moving mold temperature P2 at the start of pouring, the tank vacuum pressure P3, the mold clamping completion position P4, the chip galling pressure P5, and the cooling water flow rate P6. Therefore, the execution device 91 can predict the quality of the cast product, which may change in accordance with these input values that fluctuate with each casting of the cast product.
[0041] <Second Embodiment> The die-casting system equipped with the quality prediction device of the second embodiment will be described below with reference to the drawings. The second embodiment differs from the first embodiment in that it predicts quality mainly using the reduction width P11 of the injection speed. Components similar to those in the first embodiment are denoted by the same reference numerals and detailed descriptions are omitted.
[0042] <Overview of die casting> As shown in Figure 3, the die-casting system 110 includes a die-casting device 120 and a quality prediction device 190. The die-casting device 120 is also equipped with a speed sensor 87 that detects the injection speed, which is the speed at which the plunger 50 moves during injection by the plunger 50.
[0043] The control device 70 controls the injection speed of the plunger 50 during injection. Specifically, the control device 70 changes the target value of the injection speed of the plunger 50 from the starting position PSS, which is the stroke position when injection begins, to the ending position PSE, which is the stroke position when injection ends. The control device 70 changes the injection speed of the plunger 50 in four ranges from the starting position PSS to the ending position PSE.
[0044] As shown in Figure 4, the control device 70 controls the injection speed of the plunger 50 by changing it so that it includes a low-speed holding range AR1, an acceleration range AR2, a high-speed holding range AR3, and a deceleration range AR4 from the starting position PSS to the ending position PSE.
[0045] In the low-speed holding range AR1, the control device 70 maintains the target value of the injection velocity from the starting position PSS to the predetermined first position PS1 so as not to exceed the predetermined first velocity V1, and gradually accelerates the target value of the injection velocity from zero to the first velocity V1.
[0046] In acceleration range AR2, the control device 70 accelerates the target injection velocity from the first position PS1 to the second position PS2, which is a predetermined position closer to the end position PSE than the second position PS2, to a second velocity V2 that is faster than the first velocity V1. The acceleration range AR2 is located after the low-speed holding range AR1.
[0047] In the high-speed holding range AR3, the control device 70 maintains the target value of the injection velocity at the second velocity V2 from the second position PS2 to the third position PS3, which is predetermined to be closer to the end position PSE than the second position PS2. The high-speed holding range AR3 is located after the acceleration range AR2. The third position PS3 is the switching position from the high-speed holding range AR3 to the deceleration range AR4.
[0048] In the deceleration range AR4, the control device 70 reduces the target injection velocity from the second velocity V2 to zero from the third position PS3 to the end position PSE. The deceleration range AR4 is located after the high-speed holding range AR3.
[0049] The velocity sensor 87 detects the injection velocity at a predetermined position SP1, which is included in the low-speed holding range AR1, as the first injection velocity P7. In other words, the first specific position SP1 is located between the starting position PSS and the first position PS1.
[0050] The velocity sensor 87 detects the ejection velocity at a predetermined position SP2, which is included in the high-speed holding range AR3, as the second ejection velocity P8. The second specific position SP2 is, for example, the position where the maximum velocity is expected to occur.
[0051] The velocity sensor 87 detects the ejection velocity at a predetermined position SP3, which is included in the high-speed holding range AR3, as the third ejection velocity P9. The third specific position SP3 is closer to the third position PS3 than the second specific position SP2. The third specific position SP3 is a predetermined distance closer to the acceleration range AR2 side from the third position PS3.
[0052] The speed sensor 87 detects the ejection velocity at the fourth specific position SP4, which is a predetermined position included in the deceleration range AR4, as the fourth ejection velocity P10. The fourth specific position SP4 is defined as being closer to the end position PSE than the third specific position SP3, and also closer to the high-speed holding range AR3 by a predetermined distance from the end position PSE, but not reaching the end position PSE.
[0053] The speed sensor 87 outputs the detected first injection speed P7, second injection speed P8, third injection speed P9, and fourth injection speed P10 to the quality prediction device 190. The quality prediction device 190 acquires the first injection speed P7, the second injection speed P8, the third injection speed P9, and the fourth injection speed P10 from the speed sensor 87.
[0054] The quality prediction device 190 comprises an execution device 191 and a storage device 192. The storage device 192 stores a prediction program PR2 for predicting the quality of the cast products cast by the die-casting device 120. The execution device 191 executes the prediction program PR2, thereby allowing the quality prediction device 190 to predict the quality of the cast products cast by the die-casting device 120.
[0055] The memory device 192 stores the calculation model MD2, which is used in conjunction with the execution of the prediction program PR2. When the calculation model MD2 receives multiple input values, including multiple input values, it outputs the total volume TV of casting defects that it predicts will occur in the cast product cast by the die-casting apparatus 120 as an output value. The total volume TV of casting defects is a parameter that indicates the amount of casting defects.
[0056] The multiple input values include the fixed mold temperature P1 at the start of injection, the moving mold temperature P2 at the start of pouring, the tank vacuum pressure P3, the mold clamping completion position P4, the tip galling pressure P5, and the cooling water flow rate P6. The multiple input values further include the deceleration width P11, which is the difference between the third injection speed P9 and the fourth injection speed P10.
[0057] The calculation model MD2 is predetermined through testing and simulation. The calculation model MD2 is a machine learning model defined by machine learning using training data. The training data used in machine learning is the total volume TV of casting defects in the cast product, obtained using measuring instruments.
[0058] When the casting of the cast product by the die-casting apparatus 120 is completed, the execution device 191 starts executing the prediction program PR2. In other words, the execution device 191 starts executing the prediction program PR2 each time a cast is made.
[0059] As shown in Figure 5, when the execution device 191 starts executing the prediction program PR2, the execution device 191 first performs the process in step S21. In step S21, the execution device 191 obtains the fixed mold temperature P1 at the start of injection, the moving mold temperature P2 at the start of pouring, the tank vacuum pressure P3, the mold clamping completion position P4, the tip galling pressure P5, and the cooling water flow rate P6. In step S21, the execution device 191 further obtains the first injection speed P7, the second injection speed P8, the third injection speed P9, and the fourth injection speed P10. After that, the execution device 191 proceeds to step S22.
[0060] In step S22, the execution device 191 calculates the deceleration width P11 by subtracting the fourth injection velocity P10 from the third injection velocity P9. After that, the execution device 191 proceeds to step S23.
[0061] In step S23, the execution device 191 inputs multiple input values into the calculation model MD2. These input values include the fixed mold temperature P1 at the start of injection, the moving mold temperature P2 at the start of pouring, the tank vacuum pressure P3, the mold clamping completion position P4, the tip galling pressure P5, the cooling water flow rate P6, and the deceleration width P11. After the execution device 191 inputs the multiple inputs into the calculation model MD2, the execution device 191 proceeds to step S24.
[0062] In step S24, the execution device 191 obtains the total volume TV of the casting defects in the cast product as an output value from the calculation model MD2. A larger value for the total volume TV of the casting defects indicates a larger amount of defects. After that, the execution device 191 proceeds to step S25.
[0063] In step S25, the execution device 191 predicts the quality of the casting based on the output values output from the acquired calculation model MD2. For example, the execution device 191 calculates a quality parameter that increases as the total volume TV of casting defects increases. A smaller value for the quality parameter indicates better quality. After that, the execution device 191 completes this series of processes.
[0064] <Effects of the second embodiment> The quality prediction device 190 in the second embodiment provides the following effects in addition to the effect (1) in the first embodiment.
[0065] (2) Multiple input values further include the deceleration width P11. The detected third injection speed P9 and fourth injection speed P10 do not necessarily coincide with the target values of the injection speed of the control device 70. The difference in deceleration width P11 due to the deviation from the target value for each injection correlates with the occurrence of casting defects. Therefore, by including the deceleration width P11 in multiple input values, the quality prediction device 190 can predict the quality of the cast product more accurately.
[0066] <Example of changes> Each of the above embodiments can be implemented with the following modifications. Each of the above embodiments and the following modifications can be combined with each other to the extent that they do not contradict each other technically.
[0067] The calculation model (MD) is not limited to machine learning models using training data; it may also be a calculation formula determined in advance through testing or simulation. The storage device 192 may store the calculation model MD in the first embodiment in addition to the calculation model MD2. In the second embodiment, the execution device 191 may determine whether or not to include the deceleration width P11 in the multiple input values based on the acquired first injection velocity P7 and second injection velocity P8. When the execution device 191 determines that the deceleration width P11 should be included in the multiple input values, it may perform the processing from step S22 onwards. Alternatively, when the execution device 191 determines that the deceleration width P11 should not be included in the multiple input values, the execution device 191 may use the calculation model MD to perform the processing from step 12 onwards in the first embodiment. Regarding the determination of whether or not to include the deceleration width P11 in the multiple input values, the execution device 191 may determine that the acceleration from the first injection velocity P7 to the second injection velocity P8 is appropriate if the acceleration is greater than or equal to a predetermined acceleration width, and therefore determine to include the deceleration width P11 in the multiple input values. [Explanation of Symbols]
[0068] 10,110…Die casting system 20,120…Die casting equipment 30…Mold 31…Fixed mold 32…Moving mold 40…Sleeve 50…Plunger 52…Plunger tip 60…Vacuum tank 90,190…Quality prediction equipment 91,191…Execution equipment 92,192…Memory device C…Cavity MD,MD2…Calculation model P1…Fixed mold temperature P2…Moving mold temperature P3…Tank vacuum pressure P4…Mold clamping completion position P5…Tip galling pressure P6…Cooling water flow rate P11…Reduction width
Claims
1. A quality prediction device for predicting the quality of a cast product cast by a die-casting apparatus comprising: a fixed mold; a movable mold that is clamped to the fixed mold and partitions a cavity together with the fixed mold; a sleeve having an internal space connected to the cavity; a plunger for injecting molten metal poured into the internal space into the cavity; and a vacuum tank for reducing the pressure in the cavity, wherein the device predicts the quality of a cast product cast by the die-casting apparatus, It comprises a memory device and an execution device, The aforementioned storage device stores the calculation model, The calculation model, upon receiving a plurality of input values including the fixed mold temperature at the start of injection (the temperature of the fixed mold when the plunger starts injection), the mobile mold temperature at the start of pouring (the temperature of the mobile mold when molten metal is poured into the internal space), the tank vacuum pressure of the vacuum tank for reducing the pressure of the cavity, the clamping completion position indicating the position where the clamping operation of the mobile mold to the fixed mold is completed, the tip galling pressure (the pressure applied to the plunger tip when the plunger tip slides), and the cooling water flow rate (the flow rate of cooling water for cooling the fixed mold and the mobile mold), outputs a parameter as an output value indicating the amount of casting defects predicted to occur in the cast product. The execution device predicts the quality of the cast product based on the output value output by the calculation model. Quality prediction device.
2. The die-casting apparatus further includes a control device for controlling the injection speed of the plunger, The control device controls the ejection speed of the plunger so as the plunger moves from a starting position when it begins ejecting the plunger to an ending position when it ends the ejection, including an acceleration range for accelerating the ejection speed of the plunger, a holding range for maintaining the ejection speed of the plunger after the acceleration range, and a deceleration range for reducing the ejection speed of the plunger after the holding range. The aforementioned multiple input values further include a deceleration range, which is the difference between the ejection speed at a position that is a predetermined distance closer to the acceleration range side from the switching position between the holding range and the deceleration range, and the ejection speed at a position that is a predetermined distance closer to the holding range side from the end position. The quality prediction device according to claim 1.