Life prediction method, life prediction device, and computer program

Abstract

To provide a life prediction method that can accurately predict the life of each component constituting an industrial machine.SOLUTION: The life prediction method includes: acquiring physical quantity data indicating states of components constituting an industrial machine; determining whether or not the components are normal based on the acquired physical quantity data; and starting life prediction of the components when it is determined that the components are not normal.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present invention relates to a life prediction method, a life prediction device, and a computer program.

Background Art

[0002] An injection molding machine includes an injection device for melting and injecting a molding material and a mold clamping device. The injection device includes a heating cylinder having a nozzle at its tip and a screw rotatably arranged in the heating cylinder in the circumferential and axial directions. The screw is driven in the rotational and axial directions by a drive mechanism. The screw drive mechanism includes a ball screw that converts the rotational driving force of an injection servo motor into a driving force in the axial direction of the screw and transmits it (for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] By the way, the timing of the inspection standard of the ball screw is displayed based on the usage conditions and usage time of the injection molding machine. Further, a failure determination is made by comparing the vibration acceleration of the ball screw with a failure determination threshold value.

[0005] However, even when the inspection standard timing is reached, the degree of damage to the ball screw differs for each injection molding machine. Also, the life of the ball screw cannot be predicted by simple threshold determination. Such problems can occur similarly in all industrial machines equipped with ball screws, not limited to injection molding machines.

[0006] An object of the present disclosure is to provide a life prediction method, a life prediction device, and a computer program capable of accurately predicting the life of components constituting industrial machines. [Means for solving the problem]

[0007] A life prediction method according to one aspect of this disclosure acquires physical quantity data indicating the state of a component constituting an industrial machine, determines whether the component is normal or not based on the acquired physical quantity data, and if it is determined that the component is not normal, starts life prediction for the component.

[0008] A life prediction device according to one aspect of the present disclosure comprises an acquisition unit that acquires physical quantity data indicating the state of a component constituting an industrial machine, and a calculation unit, wherein the calculation unit determines whether the component is normal or not based on the acquired physical quantity data, and if it determines that the component is not normal, it starts predicting the life of the component.

[0009] A computer program (program product) according to one aspect of this disclosure acquires physical quantity data indicating the state of components constituting an industrial machine, determines whether the components are normal or not based on the acquired physical quantity data, and if it is determined that the components are not normal, causes the computer to execute a process to start predicting the lifespan of the components. [Effects of the Invention]

[0010] According to this disclosure, the lifespan of components that make up industrial machinery can be accurately predicted. [Brief explanation of the drawing]

[0011] [Figure 1] This is a schematic diagram showing an example of the configuration of an injection molding machine according to Embodiment 1. [Figure 2] This is a cross-sectional view showing an example of the configuration of the drive unit of an injection molding machine according to Embodiment 1. [Figure 3] This is a flowchart showing the processing procedure of the processor according to Embodiment 1. [Figure 4] This is an explanatory diagram showing the vibration acceleration signal indicating the state of a ball screw. [Figure 5A] This is an explanatory diagram showing the FFT signal intensity indicating the state of a ball screw and the method for determining a normal state. [Figure 5B] This is an explanatory diagram showing the FFT signal intensity indicating the state of a ball screw and the method for determining a normal state. [Figure 6] This is an explanatory diagram showing a ball screw degradation model and a method for predicting its lifespan. [Figure 7] This is a schematic diagram showing an example of the configuration of the control device according to Embodiment 2. [Modes for carrying out the invention]

[0012] Specific examples of an injection molding machine (life prediction device), life prediction method, and computer program according to the present invention will be described below with reference to the drawings. At least some of the embodiments described below may be arbitrarily combined. However, the present invention is not limited to these examples and is intended to be shown in the claims, with all modifications in the sense and scope equivalent to the claims being included.

[0013] (Embodiment 1) Figure 1 is a schematic diagram showing an example of the configuration of an injection molding machine 1 according to Embodiment 1. The injection molding machine 1 according to Embodiment 1 comprises a mold clamping device 2 for clamping the mold 21, an injection device 3 for plasticizing and injecting the molding material, and a control device 4. The control device 4 functions as a life prediction device according to Embodiment 1.

[0014] The clamping device 2 comprises a fixed platen 22 fixed on the bed 20, a clamping housing 23 slidably mounted on the bed 20, and a movable platen 24 that similarly slides on the bed 20. The fixed platen 22 and the clamping housing 23 are connected by multiple tie bars 25, 25, ... for example, four tie bars. The movable platen 24 is configured to slide freely between the fixed platen 22 and the clamping housing 23. A clamping mechanism 26 is provided between the clamping housing 23 and the movable platen 24.

[0015] The clamping mechanism 26 is composed of, for example, a toggle mechanism. Note that the clamping mechanism 26 may be a direct pressure type clamping mechanism, that is, it may be composed of a clamping cylinder. A fixed mold 21a and a movable mold 21b are provided on the fixed platen 22 and the movable platen 24 respectively, and when the clamping mechanism 26 is driven, the mold 21 is opened and closed.

[0016] The injection device 3 is provided on the base 30. The injection device 3 includes a heating cylinder 31 having a nozzle 31a at its tip, and a screw 32 rotatably arranged in the heating cylinder 31 in the circumferential and axial directions. A heater for melting the molding material is provided inside or on the outer circumference of the heating cylinder 31. The screw 32 is driven in the rotational and axial directions by a driving device 5 having a ball screw 51 (see FIG. 2).

[0017] Near the rear end of the heating cylinder 31, a hopper 33 for charging the molding material is provided. Further, the injection molding machine 1 includes a nozzle touch device 34 for moving the injection device 3 in the front-rear direction (left-right direction in FIG. 1). When the nozzle touch device 34 is driven, the injection device 3 advances so that the nozzle 31a of the heating cylinder 31 touches the contact portion of the fixed platen 22.

[0018] The control device 4 is a computer that controls the operations of the clamping device 2 and the injection device 3, and includes a processor (computing unit) 41, a storage unit 42, an operation unit 43, an acquisition unit 44, a display unit 45, etc. as a hardware configuration. Note that the control device 4 may be a server device connected to a network. Further, the control device 4 may be configured by a plurality of computers for distributed processing, or may be realized by a plurality of virtual machines provided in one server, or may be realized using a cloud server.

[0019] The processor 41 includes an arithmetic circuit such as a CPU (Central Processing Unit), a multi-core CPU, a GPU (Graphics Processing Unit), a GPGPU (General-purpose computing on graphics processing units), a TPU (Tensor Processing Unit), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), an NPU (Neural Processing Unit), etc., an internal storage device such as a ROM (Read Only Memory) and a RAM (Random Access Memory), an I / O terminal, a timing unit, etc. The processor 41 implements the remaining life prediction method according to Embodiment 1 by executing a computer program (program product) 42a stored in a storage unit 42 described later. Each functional unit of the control device 4 may be realized software-wise or partly or entirely hardware-wise.

[0020] The storage unit 42 is a non-volatile memory such as a hard disk, an EEPROM (Electrically Erasable Programmable ROM), or a flash memory. The storage unit 42 stores a computer program 42a for causing a computer to execute the remaining life prediction process of the ball screw 51.

[0021] The computer program 42a according to Embodiment 1 may be in a form recorded on a recording medium 6 in a computer-readable manner. The storage unit 42 stores the computer program 42a read from the recording medium 6 by a reading device. The recording medium 6 is a semiconductor memory such as a flash memory. The recording medium 6 may also be an optical disc such as a CD (Compact Disc)-ROM, a DVD (Digital Versatile Disc)-ROM, or a BD (Blu-ray (registered trademark) Disc). Further, the recording medium 6 may be a magnetic disc such as a flexible disc or a hard disk, a magneto-optical disc, etc.

[0022] Furthermore, the computer program 42a according to this embodiment 1 may be downloaded from an external server connected to the communication network and stored in the storage unit 42.

[0023] The operation unit 43 is an input device such as a touch panel, soft keys, hard keys, keyboard, or mouse.

[0024] The acquisition unit 44 performs AD conversion on the rotation angle signal output from the encoder 50d (described later) and the vibration acceleration signal output from the vibration acceleration sensor 5a to acquire rotation angle data and vibration acceleration data.

[0025] The display unit 45 is a liquid crystal panel, an organic EL display, an electronic paper display, a plasma display, etc. The display unit 45 displays various information corresponding to the image data provided by the processor 41.

[0026] The injection molding machine 1 is configured with set values ​​that define molding conditions such as the injection start time, mold resin temperature, nozzle temperature, cylinder temperature, hopper temperature, clamping force, injection speed, injection acceleration, injection peak pressure, and injection stroke.

[0027] Furthermore, the injection molding machine 1 is configured with set values ​​that define molding conditions such as the resin pressure at the cylinder tip, the seating state of the anti-reverse ring, the holding pressure switching pressure, the holding pressure switching speed, the holding pressure switching position, the holding pressure completion position, the cushion position, the metered back pressure, and the metered torque.

[0028] Furthermore, the injection molding machine 1 is configured with settings that define molding conditions such as metering completion position, screw retraction speed, cycle time, mold closing time, injection time, holding pressure time, metering time, and mold opening time. Once these settings are configured, the injection molding machine 1 operates according to those settings.

[0029] Figure 2 is a cross-sectional view showing an example of the configuration of the drive unit 5 of the injection molding machine 1 according to Embodiment 1. The drive unit 5 includes an injection servo motor 50 and a ball screw 51 for driving the screw 32 in the axial direction. The injection servo motor 50 is provided with an encoder 50d that detects the rotation angle and outputs a rotation angle signal indicating the rotation angle to the control device 4. The control device 4 controls the rotation of the injection servo motor 50 based on the rotation angle signal output from the encoder 50d. A small pulley 50a is provided on the output shaft of the injection servo motor 50.

[0030] The ball screw 51 comprises a ball screw shaft 51a and a nut 51b screwed onto the ball screw shaft 51a. The base end of the ball screw shaft 51a is rotatably supported on a first plate 52 having a hole and a bearing seat via a bearing 52a. A large pulley 50c is provided at the base end of the ball screw shaft 51a. The small pulley 50a and the large pulley 50c are connected by a timing belt 50b, and the rotational force of the small pulley 50a is transmitted to the large pulley 50c at a reduced speed, causing the ball screw shaft 51a to rotate.

[0031] Hereinafter, the direction toward the base end of the ball screw shaft 51a (right side in Figures 1 and 2) will be referred to as the retraction direction, and the direction toward the opposite end (left side in Figures 1 and 2) will be referred to as the advance direction. The advance and retraction directions together will be referred to as the forward and backward directions. When the injection servo motor 50 is driven and the large pulley 50c and ball screw shaft 51a rotate, the nut 51b moves in the forward and retraction directions according to the direction of rotation.

[0032] A load cell 53 is provided on the forward-facing side of the nut 51b, and the forward-facing side of the load cell 53 is fixed to the second plate 54. Multiple through holes are formed in the second plate 54, and a guide shaft 55 is inserted through these through holes. The second plate 54 moves in the forward and backward directions guided by the guide shaft 55. A third plate 56 is provided on the forward-facing side of the first plate 52, and one end and the other end of the guide shaft 55 are supported by the first plate 52 and the third plate 56, respectively. When the ball screw shaft 51a rotates, the nut 51b, load cell 53, and second plate 54 move integrally in the forward and backward directions along the guide shaft 55.

[0033] The second plate 54 is provided with a hole and a bearing seat, and the output shaft 58 is supported in the hole via a bearing 54a. A plasticizing pulley 57 is provided on the output shaft 58. The plasticizing pulley 57 is connected to a pulley attached to a screw rotation motor (not shown) via a timing belt (not shown). The rotation center of the output shaft 58 coincides with the rotation center of the ball screw shaft 51a. The output shaft 58 has a recess formed in it into which the tip of the ball screw shaft 51a enters when the nut 51b moves back and forth.

[0034] Furthermore, one end of the screw 32 is fixed to the output shaft 58 so that its central axis coincides with the output shaft 58. The third plate 56 has a through hole through which the screw 32 is inserted. One end of the heating cylinder 31 is fixed to the third plate 56 so that the screw 32, inserted through the through hole in the third plate 56, can move axially within the heating cylinder 31.

[0035] Furthermore, a vibration acceleration sensor 5a is attached to the nut 51b of the ball screw 51 to detect vibrations of the nut 51b. The vibration acceleration sensor 5a outputs the detected vibration acceleration signal to the control device 4. The processor 41 of the control device 4 acquires the vibration acceleration signal output from the vibration acceleration sensor 5a by performing an A / D conversion to acquire vibration acceleration data in the acquisition unit 44. Based on the vibration acceleration data, the control device 4 can predict the lifespan of the ball screw 51.

[0036] The molding process cycle is outlined below, and the control device 4 performs life prediction processing for the ball screw 51 in the repeated molding process cycle. During injection molding, the well-known mold closing process, mold clamping process, injection unit advance process, injection process, metering process, injection unit retraction process, mold opening process, and ejection process are performed in sequence.

[0037] <Method for predicting lifespan> A method for predicting the lifespan of a ball screw 51 installed in an injection molding machine 1 configured as described above will now be explained. When foreign matter gets caught between the ball screw shaft 51a and the nut 51b of the ball screw 51, tiny scratches are made on the screw surface, and as the ball passes over the scratched area, the screw surface peels off from that area. Furthermore, when the ball steps on the peeled-off foreign matter, new scratches are created, and wear progresses at an accelerating rate.

[0038] Therefore, predicting the lifespan of the ball screw 51 by monitoring its vibrations is extremely difficult until the first damage affecting its lifespan occurs. Furthermore, if lifespan prediction is performed using accumulated vibration acceleration data of the ball screw 51 without considering the occurrence of events that affect its lifespan, the accuracy of the lifespan prediction will decrease.

[0039] In injection molding machine 1, resin containing glass fibers or metal powder may be used, and the glass fibers or metal powder can easily get caught in the ball screw 51, creating an environment where damage is likely to occur. In other words, the ball screw 51 of injection molding machine 1 is in an environment where it is difficult to predict its lifespan.

[0040] The control device (life prediction device) 4 of this disclosure predicts the life of the ball screw 51 by a two-stage determination algorithm that uses a normal state determination algorithm and a life prediction algorithm.

[0041] The normal state determination algorithm is an algorithm that determines whether the ball screw 51 is normal or not. In other words, the normal state determination algorithm detects the occurrence of an event that causes the wear of the ball screw 51 to progress at an accelerating rate. The occurrence of this event causes the deterioration of the ball screw 51 to begin. In this embodiment 1, an example is described in which the ball screw 51 is determined to be normal or not by comparing the magnitude of the frequency component of the vibration acceleration data in a specific frequency band with a threshold value. Details will be described later.

[0042] The life prediction algorithm predicts the life of the ball screw 51 based on vibration acceleration data indicating the vibration of the ball screw 51 when the above-mentioned event occurs. In this embodiment 1, an example of predicting the life by RUL (remaining useful life) estimation using a degradation model of a state indicator (index value) correlated with the life of the ball screw 51 is described. Details will be described later.

[0043] The two-stage judgment algorithm uses a normal state judgment algorithm in the initial stage to detect the start of deterioration of the ball screw 51 (damage to the nut 51b). Once deterioration of the ball screw 51 begins, the wear and deterioration of the ball screw 51 progresses at an accelerating rate. As the vibration state of the ball screw 51 changes during this wear and deterioration process, the lifespan prediction algorithm can be used to more accurately predict the lifespan of the ball screw 51.

[0044] Figure 3 is a flowchart showing the processing procedure of the processor 41 according to Embodiment 1. Steps S111 to S114 correspond to the normal state determination algorithm. Steps S115 to S119 correspond to the life prediction algorithm.

[0045] The processor 41 acquires vibration acceleration data indicating the state of the ball screw 51 from the vibration acceleration sensor 5a via the acquisition unit 44 (step S111).

[0046] Figure 4 is an explanatory diagram showing the vibration acceleration signal indicating the state of the ball screw 51. In the graph shown in Figure 4, the horizontal axis represents time, and the vertical axis represents the vibration acceleration signal output by the vibration acceleration sensor 5a. The processor 41 acquires the vibration acceleration signal by performing an A / D conversion to vibration acceleration data in the acquisition unit 44.

[0047] Next, the processor 41 stores the acquired vibration acceleration data and time data indicating the time when the vibration acceleration data was acquired in the storage unit 42 (step S112). The time data is obtained from the timing unit.

[0048] Next, the processor 41 performs a Fast Fourier Transform on the vibration acceleration data and time data stored in the memory unit 42 (step S113). The Fast Fourier Transform yields the frequency components of the vibration acceleration data. Hereinafter, the magnitude of the frequency components will be referred to as the FFT (Fast Fourier Transform) signal strength.

[0049] Figures 5A and 5B are explanatory diagrams showing the FFT signal intensity indicating the state of the ball screw 51 and the method for determining the normal state. In the graphs shown in Figures 5A and 5B, the horizontal axis represents frequency and the vertical axis represents FFT signal intensity. Figure 5A shows the FFT signal intensity obtained from a normal ball screw 51. Figure 5B shows the FFT signal intensity obtained from a non-normal ball screw 51. In other words, Figure 5B shows the FFT signal intensity obtained when an event occurs that causes the wear of the ball screw 51 to progress at an accelerating rate.

[0050] If the nut 51b is damaged, the FFT signal intensity values ​​in the specific frequency band indicated by the rectangular frame in Figures 5A and 5B increase overall. In Figures 5A and 5B, the dashed lines indicate thresholds for determining whether the ball screw 51 is normal or abnormal. The specific frequency band and threshold can be obtained experimentally using a normal ball screw 51. The memory unit 42 stores the specific frequency band and threshold.

[0051] After completing the process in step S113, the processor 41 determines whether the ball screw 51 is normal or not by comparing the FFT signal strength in a specific frequency band with a threshold (step S114). For example, the processor 41 determines that the ball screw 51 is normal if the average value of the FFT signal strength in a specific frequency band is less than the threshold. The processor 41 determines that the ball screw 51 is not normal if the average value of the FFT signal strength in a specific frequency band is greater than or equal to the threshold.

[0052] If the ball screw 51 is determined to be normal (step S114: YES), the processor 41 returns to step S111 and continues the normal state determination process to determine whether or not the ball screw 51 is normal.

[0053] If the ball screw 51 is determined to be abnormal (step S114: NO), the processor 41 acquires vibration acceleration data from the vibration acceleration sensor 5a in order to predict the lifespan of the ball screw 51 (step S115). The processor 41 stores the acquired vibration acceleration data and time data indicating the time when the vibration acceleration data was acquired in the storage unit 42 (step S116).

[0054] Next, the processor 41 calculates a state indicator (index) correlated with the lifespan of the ball screw 51 based on the vibration acceleration data stored in the memory unit 42 (step S117). The state indicator is, for example, the peak value of the vibration acceleration data. The vibration acceleration data used for lifespan prediction is the vibration acceleration data accumulated after the point in time when the ball screw 51 is determined to be abnormal.

[0055] The processor 41 calculates a degradation model for estimating the time-dependent change of the state indicator, which is correlated with the lifespan of the ball screw 51, based on the state indicator and time data based on the vibration acceleration data stored in the memory unit 42 (step S118). The degradation model is a function that shows the relationship between usage time and the state indicator. The degradation model can be obtained by the maximum likelihood estimation method.

[0056] The function related to the degradation model can be expressed, for example, by equation (1) below. P = a × e bt +c…(1) however, P: Status indicator t: Usage time a, b, c: coefficients

[0057] The processor 41 finds the degradation model by calculating the coefficients a, b, and c that minimize the mean squared error. The confidence interval of the degradation model can be found in the same way.

[0058] Figure 6 is an explanatory diagram showing the degradation model and life prediction method for the ball screw 51. In the graph shown in Figure 6, the horizontal axis represents the operating time of the ball screw 51, and the vertical axis represents the state indicator of the ball screw 51. The irregularly increasing line shows the change in the state indicator over time. The smooth curve is the degradation model of the ball screw 51, and the dashed line shows the confidence interval of the degradation model. The thick line is the failure detection threshold. The failure detection threshold is the value of the state indicator when the ball screw 51 fails.

[0059] The lifespan of the ball screw 51 can be predicted by using a degradation model to determine the usage time at which the status indicator exceeds the failure threshold. The failure threshold can be experimentally obtained using a damaged ball screw 51. The memory unit 42 stores the failure threshold.

[0060] Next, the processor 41 predicts the lifespan of the ball screw 51 using a degradation model (step S119). For example, the processor 41 can estimate the failure time as the point at which the state indicator calculated using the degradation model reaches a failure threshold. The processor 41 can determine the remaining service life by subtracting the current usage time from the failure time.

[0061] Next, the processor 41 outputs the life prediction result (step S120). For example, the processor 41 displays the failure time or remaining service life on the display unit 45. The processor 41 may also display a graph on the display unit 45 as shown in Figure 6.

[0062] Next, the processor 41 performs a Fast Fourier Transform on the vibration acceleration data (step S121), similar to steps S113 and S114, and determines whether the ball screw 51 is functioning correctly (step S122). In other words, the processor 41 determines whether the ball screw 51 is functioning correctly even while life prediction is in progress.

[0063] If the ball screw 51 is determined to be abnormal (step S122: NO), the processor 41 returns to step S115 and continues the life prediction process.

[0064] If the ball screw 51 is determined to be normal (step S122: YES), the processor 41 stops the life prediction process (step S123), stops outputting the life prediction result (step S124), and returns the process to step S111.

[0065] As described above, the injection molding machine 1 according to this embodiment 1 can accurately predict the lifespan of the ball screw 51 of the injection molding machine 1. Specifically, the lifespan of the ball screw 51 can be accurately predicted by a two-stage determination algorithm that uses a normal state determination algorithm and a lifespan prediction algorithm.

[0066] If the ball screw 51 is in a normal state, that is, if it is difficult to accurately predict the lifespan of the ball screw 51, the processor 41 does not perform a lifespan prediction and does not display the lifespan prediction result on the display unit 45. The operator of the injection molding machine 1 can recognize that the ball screw 51 is not in a state where deterioration is progressing. The operator can use the injection molding machine 1 with confidence until the ball screw 51 deviates from its normal state.

[0067] When the ball screw 51 deviates from its normal state, the processor 41 starts predicting the lifespan of the ball screw 51 using a lifespan prediction algorithm and displays the lifespan prediction result on the display unit 45. The operator can recognize that the ball screw 51 is deteriorating. The operator can then determine the lifespan of the ball screw 51 from the lifespan prediction result displayed on the display unit 45.

[0068] The processor 41 starts predicting the lifespan of the ball screw 51 when wear on the ball screw 51 is progressing and the condition indicator may change, thus enabling it to predict the lifespan of the ball screw 51 more accurately. In other words, compared to when the lifespan prediction starts from when the ball screw 51 is in a normal state, the processor 41 can predict the lifespan of the ball screw 51 with greater accuracy.

[0069] Furthermore, according to the processing in steps S121 to S124, even if the life prediction process is started due to a misjudgment, the life prediction can be stopped and the process can return to monitoring the state of the ball screw 51. This prevents the life prediction process from continuing even though the ball screw 51 is in a normal state, and prevents the output of a life prediction result with low prediction accuracy or an incorrect result.

[0070] If the ball screw 51 is determined to be normal during the life prediction process, the processor 41 temporarily stops the life prediction process and the output process of the life prediction results. If the ball screw 51 is again determined to be abnormal, the processor 41 restarts the life prediction process from the beginning, thereby preventing a decrease in the accuracy of the life prediction. Even if the condition of the ball screw 51 is misjudged, the life of the ball screw 51 can still be accurately predicted.

[0071] Furthermore, if the ball screw 51 is replaced, the processor 41 can return to the ball screw 51 status monitoring state without performing any special processing.

[0072] The normal state determination algorithm described in this embodiment 1 is just one example; the control device 4 may be configured to determine whether the ball screw 51 is normal or not using the peak value of the FFT signal intensity or other statistical values.

[0073] Furthermore, the processor 41 may use a current sensor to acquire drive current data of the injection servo motor 50 and determine whether the ball screw 51 is functioning correctly. The processor 41 makes the above determination by comparing the average value, peak value, and other statistical values ​​of the FFT signal intensity obtained by fast Fourier transforming the drive current data with a threshold value.

[0074] The processor 41 may use a torque sensor to acquire torque data from the injection servo motor 50 and determine whether the ball screw 51 is functioning correctly. The processor 41 makes the above determination by comparing the average value, peak value, and other statistical values ​​of the FFT signal intensity obtained by fast Fourier transforming the torque data with a threshold value.

[0075] The processor 41 may determine whether the ball screw 51 is functioning correctly by comparing the average value, peak value, and other statistical values ​​of the FFT signal intensity obtained by performing a fast Fourier transform on the drive control amount data with a threshold value.

[0076] The processor 41 may use a microphone to acquire drive sound data of the injection servo motor 50 and determine whether the ball screw 51 is functioning correctly. The processor 41 makes the above determination by comparing the average value, peak value, and other statistical values ​​of the FFT signal intensity obtained by fast Fourier transforming the drive sound data with a threshold.

[0077] When vibration acceleration data indicating the vibration of the ball screw 51 is input, a state determination learning model that has been trained to output data indicating whether or not the ball screw 51 is normal may be used to determine whether or not the ball screw 51 is normal.

[0078] The control device 4 may be configured to determine whether the ball screw 51 is functioning correctly or not by using a state determination learning model that has been trained to output data indicating whether the ball screw 51 is functioning correctly or not when drive current data for the injection servo motor 50 is input.

[0079] The control device 4 may be configured to determine whether the ball screw 51 is functioning correctly or not by using a state determination learning model that has been trained to output data indicating whether the ball screw 51 is functioning correctly or not when torque data from the injection servo motor 50 is input.

[0080] The control device 4 may be configured to determine whether the ball screw 51 is functioning correctly or not by using a state determination learning model that has been trained to output data indicating whether the ball screw 51 is functioning correctly or not when drive control amount data for the injection servo motor 50 is input.

[0081] The control device 4 may be configured to determine whether the ball screw 51 is functioning correctly or not by using a state determination learning model that has been trained to output data indicating whether the ball screw 51 is functioning correctly or not when drive sound data of the ball screw 51 is input.

[0082] Multiple types of normal state determination algorithms described above may be prepared, and the ball screw 51 may be selectively used to determine whether it is normal or not. For example, the processor 41 may perform a determination using multiple types of normal state determination algorithms, and if a predetermined number of normal state determination algorithms determine that it is not normal, it may be determined that the ball screw 51 is abnormal. Furthermore, the ball screw 51 may be determined to be normal or not using a weighted average of the determination results obtained using the multiple types of normal state determination algorithms described above.

[0083] The life prediction algorithm described in the embodiment is just one example, and a degradation model may be created using drive current data, drive control amount data, torque data, and drive sound data obtained from the ball screw 51 of the injection servo motor 50. The processor 41 uses this degradation model to predict the life of the ball screw 51.

[0084] The peak value of vibration acceleration data is one example of a state indicator; other values ​​based on vibration acceleration data may also be used as state indicators. The average value, peak value, or other statistics of the FFT signal intensity of vibration acceleration data in a specific frequency band may also be used as a state indicator.

[0085] Similar to the normal state determination algorithm, multiple types of life prediction algorithms may be combined to predict the lifespan of the ball screw 51.

[0086] Furthermore, although this embodiment 1 describes an example of predicting the lifespan of the ball screw 51, it may also be configured to predict the lifespan of other components constituting the clamping device 2 and the injection device 3.

[0087] Furthermore, the system may be configured to acquire physical quantity data indicating the state of components of industrial machinery other than the injection molding machine 1, and to predict the lifespan of those components. Note that industrial machinery includes extruders and any other equipment used in the industrial field.

[0088] (Embodiment 2) The injection molding machine 1 according to Embodiment 2 differs from Embodiment 1 in that it predicts the lifespan of the ball screw 51 using machine learning. Other processing details, aside from the lifespan prediction algorithm, are the same as those in Embodiment 1. Since the other components of the injection molding machine 1 are the same as those in Embodiment 1, the same reference numerals are used for the same parts, and detailed explanations are omitted.

[0089] Figure 7 is a schematic diagram showing an example configuration of the control device 4 according to Embodiment 2. The processor 41 of the control device 4 according to Embodiment 2 includes a learning processing unit 41a as a functional unit. The storage unit 42 stores a learning model 42b for predicting the remaining service life of the ball screw 51. The learning processing unit 41a may be implemented in software or in hardware. Furthermore, a part of the learning processing unit 41a may be implemented in hardware.

[0090] The processor 41 reads the learning model 42b and the computer program 42a from the memory unit 42 and executes them.

[0091] The learning model 42b is a neural network that outputs the remaining lifespan when given an FFT signal intensity distribution based on vibration acceleration data as input. The learning model 42b comprises an input layer, a hidden layer, and an output layer. The input layer has multiple nodes to which the FFT signal intensity distribution is input. The hidden layer comprises multiple hidden layers, each having multiple nodes, with the nodes of the input-side hidden layers connected to the nodes of the input layer. The output layer has nodes to output the remaining lifespan. Each node of the output layer is connected to the nodes of the output-side hidden layers.

[0092] The distribution of FFT signal intensity is, for example, image data, and the learning model 42b is composed of, for example, a convolutional neural network (CNN).

[0093] The learning model 42b can be generated by machine learning using training data that includes the FFT signal intensity distribution based on vibration acceleration data and the remaining lifespan. For example, the learning processing unit 41a of the processor 41 optimizes the weight coefficients of the learning model 42b using methods such as backpropagation and gradient descent with respect to the training data, thereby enabling machine learning of the learning model 42b.

[0094] The processor 41 can calculate the remaining service life of the ball screw 51 by inputting the FFT signal intensity distribution detected at this time into the learning model 42b.

[0095] According to the injection molding machine 1 of this second embodiment, the lifespan of the ball screw 51 can be predicted using a two-stage determination algorithm, similar to the first embodiment.

[0096] In this embodiment 2, a CNN learning model 42b was described, but the system may also be configured to predict failure time and remaining service life using other known machine learning models such as neural networks other than CNNs, SVMs (Support Vector Machines), or Bayesian networks.

[0097] Furthermore, although this second embodiment describes an example in which lifespan prediction is performed using a machine learning model, the normal state determination process for the ball screw 51 may also be performed using a machine learning model. [Explanation of Symbols]

[0098] 1 injection molding machine 2 Mold clamping device 3 Injection device 4. Control device (life prediction device) 5. Drive unit 5a Vibration acceleration sensor 6. Recording media 51 Ball screw 41 processors

Claims

1. We acquire physical quantity data that indicates the state of the components that make up industrial machinery. Based on the acquired physical quantity data, it is determined whether the part is functioning correctly. If it is determined that the component is not functioning correctly, the lifespan prediction of the component will be initiated. Even after the start of the life prediction, it is determined whether the component is functioning normally. If it is determined that the component is normal, the lifespan prediction for the component is stopped. Methods for predicting lifespan.

2. The acquired physical quantity data and time data indicating the time of acquisition of the physical quantity data are stored in association with each other. Based on the acquired physical quantity data and time data, a function is calculated to estimate the change over time of an index value correlated with the lifespan of the component. The calculated function is used to predict the lifespan of the component. The life prediction method according to claim 1.

3. When the life prediction is initiated, the output of the life prediction results is started. If the result is determined to be normal, the output of the lifespan prediction result will be stopped. A method for predicting lifespan according to claim 1 or claim 2.

4. If, after discontinuing the lifespan prediction, it is determined again that the component is not functioning correctly, the lifespan prediction for the component will be restarted from the point at which it was determined to be functioning incorrectly again. A method for predicting lifespan according to any one of claims 1 to 3.

5. The aforementioned industrial machine is a molding machine having a ball screw, Predicting the lifespan of the ball screw A method for predicting lifespan according to any one of claims 1 to 4.

6. The aforementioned physical quantity data is data indicating the vibration acceleration of the ball screw, the current of the motor driving the ball screw, or the driving sound. The life prediction method according to claim 5.

7. Obtain physical quantity data indicating the state of the components constituting the industrial machine, Based on the acquired physical quantity data, it is determined whether the part is functioning correctly. If it is determined that the component is not functioning correctly, the lifespan prediction of the component will be initiated. The physical quantity data acquired after the start of life prediction and the time data indicating the time of acquisition of said physical quantity data are stored in association with each other. Based on the physical quantity data and time data acquired after the start of life prediction, a function is calculated to estimate the change over time of an index value correlated with the lifespan of the component after the start of life prediction. The calculated function is used to predict the lifespan of the component. Methods for predicting lifespan.

8. An acquisition unit that acquires physical quantity data indicating the state of the components that make up industrial machinery, Calculation unit and Equipped with, The aforementioned arithmetic unit, Based on the acquired physical quantity data, it is determined whether the part is functioning correctly. If it is determined that the component is not functioning correctly, the lifespan prediction of the component will be initiated. Even after the start of the life prediction, it is determined whether the component is functioning normally. If it is determined that the component is normal, the lifespan prediction for the component is stopped. Life prediction device.

9. An acquisition unit that acquires physical quantity data indicating the state of components constituting an industrial machine, Calculation unit and Equipped with, The aforementioned arithmetic unit, Based on the acquired physical quantity data, it is determined whether the part is functioning correctly. If it is determined that the component is not functioning correctly, the lifespan prediction of the component will be initiated. The physical quantity data acquired after the start of life prediction and the time data indicating the time of acquisition of said physical quantity data are stored in association with each other. Based on the physical quantity data and time data acquired after the start of life prediction, a function is calculated to estimate the change over time of an index value correlated with the lifespan of the component after the start of life prediction. The calculated function is used to predict the lifespan of the component. Life prediction device.

10. We acquire physical quantity data that indicates the state of the components that make up industrial machinery. Based on the acquired physical quantity data, it is determined whether the part is functioning correctly. If it is determined that the component is not functioning correctly, the lifespan prediction of the component will be initiated. Even after the start of the life prediction, it is determined whether the component is functioning normally. If it is determined that the component is normal, the lifespan prediction for the component is stopped. A computer program that causes a computer to perform a process.

11. Obtain physical quantity data indicating the state of the components constituting an industrial machine, Based on the acquired physical quantity data, it is determined whether the part is functioning correctly. If it is determined that the component is not functioning correctly, the lifespan prediction of the component will be initiated. The physical quantity data acquired after the start of life prediction and the time data indicating the time of acquisition of said physical quantity data are stored in association with each other. Based on the physical quantity data and time data acquired after the start of life prediction, a function is calculated to estimate the change over time of an index value correlated with the lifespan of the component after the start of life prediction. The calculated function is used to predict the lifespan of the component. A computer program that causes a computer to perform a process.

Images