Control device, control method, and control program
The control device for servo press machines uses load detection and machine learning to dynamically adjust stop times, addressing the balance between accuracy and efficiency by ensuring accurate finishing and reducing cycle time.
Patent Information
- Application Number
- JP2021034801
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-03-04
- Publication Date
- 2024-11-13
- Estimated Expiration
- 2041-03-04
AI Technical Summary
Servo press machines face a challenge in balancing finishing accuracy with cycle time, as increasing bottom dead center stop time improves accuracy but lengthens the machining cycle, reducing production efficiency.
A control device and method that includes a servo motor, position detection, load detection, and a control unit to dynamically adjust the bottom dead center stop time based on load convergence, using machine learning to predict load conditions and adjust operations accordingly.
Ensures finishing accuracy while shortening the cycle time by dynamically changing the bottom dead center stop time, improving productivity and reducing the negative effects of material variations.
Smart Images

Figure 0007581974000001 
Figure 0007581974000002 
Figure 0007581974000003
Abstract
Description
[Technical field]
[0001] The present disclosure relates to a control device, a control method, and a control program for controlling a servo press machine. [Background technology]
[0002] In recent years, in the field of press systems, servo press machines that perform press working on a workpiece (material) using a press tool as a die driven by a servo motor via a slider have become increasingly popular. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] JP 2011-098350 A Summary of the Invention [Problem to be solved by the invention]
[0004] In a servo press, the slider moves up and down, and the load on the material can be the greatest at the bottom dead center, which is the lowest point. In a servo press, the deformation of the workpiece is completed when the slider stops at the bottom dead center (bottom dead center stop time). Therefore, by extending the bottom dead center stop time, it is possible to improve the finishing accuracy of the product.
[0005] On the other hand, an increase in the bottom dead center time also leads to a longer cycle time for processing, resulting in a decrease in production efficiency. For this reason, servo presses are required to ensure both the accuracy of the finished product and the shortened cycle time.
[0006] The present disclosure has been made in consideration of the above-mentioned problems, and aims to provide a control device, control method, and control program that can ensure both the finishing accuracy of products produced by a servo press and the shortening of cycle time. [Means for solving the problem]
[0007] In order to solve the above-mentioned problems, the present disclosure employs the following configuration.
[0008] A control device according to one aspect of the present disclosure is a servo press machine that performs press processing on a material by moving a slider in an up and down direction, the control device controlling the servo press machine including a servo motor that drives the slider, a position detection unit that detects the position of the slider, and a load detection unit that detects the load acting on the material, and includes a control unit that controls the servo motor using a position detection result by the position detection unit and a load detection result by the load detection unit, and the control unit performs a descending operation that descends the slider toward a bottom dead center position, which is the lowest position of the slider, a stopping operation that stops the slider at the bottom dead center position, and an ascending operation that raises the slider from the bottom dead center position, as a series of step operations in the press processing, and is configured to determine whether or not the load acting on the material has converged based on the load detection result during the stopping operation, and to perform the ascending operation when it is determined that the load acting on the material has converged.
[0009] According to the above configuration, it is possible to ensure the finishing precision of the product produced by the servo press machine while also shortening the cycle time.
[0010] In the control device relating to the above-mentioned aspect, the control unit may determine that the load acting on the material is in a converged state when the difference between the load detection result newly acquired from the load detection unit and the average value of the load detection result within a predetermined period of time is less than a predetermined first threshold value.
[0011] According to the above configuration, it is possible to reliably improve the accuracy of determining whether or not the load is in a converged state.
[0012] The control device relating to the above-mentioned aspect further includes a predictor that acquires the load detection result from the load detection unit, calculates a load prediction value which is a prediction value of the load acting on the material after a predetermined time from time series data of the acquired load detection result, and outputs the load prediction value to the control unit, and the control unit may use the load prediction value during the stopping operation to determine whether or not the load acting on the material is in a converged state.
[0013] According to the above configuration, by utilizing predictions using machine learning, it is possible to more quickly determine whether the load is in a converged state, and it is possible to perform convergence determination at an early stage.
[0014] In the control device according to the above aspect, the control unit may determine that the load acting on the material is in a converged state when a difference between the load prediction value newly acquired from the predictor and an average value of the load prediction value within a predetermined period of time in the past is less than a preset second threshold value.
[0015] According to the above configuration, it is possible to reliably improve the accuracy of determining whether or not the load is in a converged state.
[0016] In the control device relating to the above-mentioned aspect, the predictor may have a learning model generated by performing machine learning using as teacher data a pair of data within a first period and data from the end of the first period among the time series data of the load detection result during the stopping operation, as training data, and generating a learning model that uses data within the first period as input and outputs data after the specified time from the end of the first period as the load prediction value.
[0017] According to the above configuration, a learning model that is machine-learned in the predictor is used, so that the bottom dead center stop time of the servo press can be dynamically changed more appropriately.
[0018] In the control device according to the aforementioned aspect, the load detection unit may be a load detection unit that detects a torque of the servo motor to thereby detect a load acting on the material.
[0019] According to the above configuration, the load acting on the material can be detected by using the torque of the servo motor without providing a separate sensor.
[0020] In the control device relating to the above-mentioned aspect, the load detection unit may be a load detection unit that includes a strain gauge provided on a punch that transmits at least a portion of the load applied by the slider to the material, and detects the load acting on the material by detecting the amount of strain using the strain gauge.
[0021] According to the above-described configuration, by using the amount of strain detected by the strain gauge, the load acting on the material can be detected more directly.
[0022] Moreover, a control method according to one aspect of the present disclosure is a control method for controlling a servo press machine that performs press processing on a material by moving a slider in an up and down direction, the servo press machine being equipped with a servo motor that drives the slider, a position detection unit that detects the position of the slider, and a load detection unit that detects a load acting on the material, the control method including a lowering operation step that lowers the slider toward a bottom dead center position that is the lowest position of the slider, a stopping operation step that stops the slider at the bottom dead center position, and an ascending operation step that raises the slider from the bottom dead center position, and during the stopping operation step, a substep of determining whether or not the load acting on the material is in a converged state based on the load detection result is repeatedly executed, and when it is determined that the load acting on the material is in a converged state, transition is made to the ascending operation step.
[0023] According to the above configuration, it is possible to ensure the finishing precision of the product produced by the servo press machine while also shortening the cycle time.
[0024] A control program according to an aspect of the present disclosure is a control program for causing a computer to function as a control device, and causes the computer to function as the control unit.
[0025] According to the above configuration, it is possible to ensure the finishing precision of the product produced by the servo press machine while also shortening the cycle time. Effect of the Invention
[0026] According to the present disclosure, it is possible to provide a control device, a control method, and a control program that can ensure the finishing accuracy of products in a servo press machine while shortening the cycle time. [Brief description of the drawings]
[0027] [Figure 1] 1 is a block diagram showing a configuration example of a control device and a servo press according to a first embodiment of the present disclosure. FIG. [Diagram 2] FIG. 2 is a functional configuration diagram showing a control system of a servo press machine by the above control device. [Diagram 3] 3A to 3C are diagrams illustrating the basic operation of the servo press machine. [Figure 4] 2 is a diagram illustrating a specific example of the configuration of a predictor shown in FIG. 1. FIG. [Diagram 5] 4 is a graph showing a specific example of changes in the load detection result in the servo press machine. [Figure 6] FIG. 4 is a diagram illustrating an example of the operation of a predictor included in the control device. [Figure 7] 4 is a diagram illustrating an example of the operation of a control unit included in the control device. FIG. [Figure 8] FIG. 13 is a diagram showing a specific formula used for determining convergence in the control unit. [Figure 9] 4 is a flowchart illustrating an example of the operation of the control device. [Figure 10] 10 is a flowchart illustrating another example of the operation of the control device. [Figure 11]FIG. 4 is an explanatory diagram for explaining a specific example of the effect of the control device. [Figure 12] FIG. 11 is a block diagram showing an example configuration of a control device and a servo press according to a second embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0028] [Embodiment 1] §1 Examples of application First, an example of a situation to which the present disclosure is applied will be described with reference to Fig. 1 to Fig. 3. Fig. 1 is a block diagram showing an example of a configuration of a control device and a servo press according to a first embodiment of the present disclosure. Fig. 2 is a functional configuration diagram showing a control system of the servo press by the control device. Fig. 3 is a diagram explaining a basic operation of the servo press.
[0029] 1 and 2, a control device 1 is used, for example, in a manufacturing site, and controls a servo press machine 10. The control device 1 is realized, for example, by a PLC (Programmable Logic Controller) or a servo driver.
[0030] The control device 1 is connected to one or more servo press machines 10. As a result, the control device 1 and the servo press machines 10 configure a press system that performs press processing on a material Z to produce a product P.
[0031] The servo press 10 is a press device whose power source is a servo motor 12 that drives a slider 11. Specifically, in the servo press 10, a rotational motion of the servo motor 12 is converted into linear motion by an actuator (not shown). The servo press 10 operates the slider 11 in a predetermined vertical direction to perform press working on the material Z that is in contact with a press tool (not shown) attached to the slider 11.
[0032] Furthermore, in the servo press 10, a series of step operations in the press process include a lowering operation, a stopping operation, and an ascending operation of the slider 11. Specifically, in the lowering operation, as shown from time T1 to time T2 in Fig. 3, the slider 11 is lowered, for example, from a top dead center position that is the highest point position in the vertical direction to a bottom dead center position that is the lowest point position in the vertical direction.
[0033] In the downward movement of the slider 11, the downward speed is divided into two stages as shown by arrows A and B in Fig. 3. In other words, as the slider 11 approaches the bottom dead center position, the downward speed is decelerated. This makes it possible to prevent the slider 11 from being unable to stop accurately at the bottom dead center position due to the influence of inertia, etc., and suppresses overload (overshoot) on the material Z, thereby significantly suppressing the occurrence of a decrease in the finishing accuracy of the product P.
[0034] In addition, in the stopping operation, the slider 11 stops at the bottom dead center position, as shown from time T2 to time T3 in Fig. 3. In addition, in the servo press 10, the finishing precision of the product P can be improved by appropriately setting the bottom dead center stopping time between time T2 and time T3 in accordance with the material and thickness of the material Z. Note that the maximum value of this bottom dead center stopping time may be determined, for example, based on a test operation by an operator.
[0035] During the upward movement, as shown from time T3 to time T4 in FIG. 3, the slider 11 ascends from the bottom dead center position to the top dead center position at a constant ascending speed (indicated by arrow C in FIG. 3), for example.
[0036] 1 and 2, the control device 1 has a function of collecting and machine learning data related to the operation of the servo press 10. The control device 1 acquires information from the servo press 10, such as the position detection result of the slider 11 from the position detection unit 13, the speed detection result of the servo motor 12 from the speed detection unit 14, and the load detection result acting on the material Z from the load detection unit 15.
[0037] During the stopping operation of the slider 11, the control unit 5 determines whether or not the load acting on the material Z has converged, using the load detection result from the load detection unit 15. Then, when the control unit 5 determines that the load acting on the material Z has converged, it causes the slider 11 to perform an upward movement from the stopping operation.
[0038] Therefore, according to this embodiment, the control device 1 can appropriately change the bottom dead center stop time in the above stopping operation, and can dynamically change the bottom dead center stop time of the slider 11 in the servo press 10 according to the load acting on the material Z. As a result, the control device 1 can ensure both the finishing accuracy of the product P in the servo press 10 and the shortening of the cycle time.
[0039] §2 Configuration Example [Embodiment 1] An embodiment of the present disclosure will be described in detail below. First, a servo press 10 that is an object to be controlled by the control device 1 of the present embodiment will be described.
[0040] <Configuration of servo press machine 10> As shown in FIGS. 1 and 2, the servo press 10 includes a slider 11, a servo motor 12, a position detection unit 13, a speed detection unit 14, and a load detection unit 15.
[0041] The slider 11 performs the above-mentioned lowering and stopping motions on the material Z in contact with the press tool by a driving force corresponding to the rotational motion of the servo motor 12, thereby performing press working on the material Z to produce the product P. The servo motor 12 performs the rotational motion according to an instruction from the control device 1.
[0042] The position detection unit 13 is a detection unit that detects the position of the slider 11, and includes, for example, a position sensor such as an optical encoder. The position detection unit 13 outputs the detection result of the position sensor to the control device 1 as the position detection result of the slider 11.
[0043] The speed detection unit 14 is a detection unit that detects the rotation speed of the servo motor 12, and includes, for example, a speed sensor using an optical encoder. The speed detection unit 14 outputs the detection result of the speed sensor to the control device 1 as the speed detection result of the servo motor 12.
[0044] The load detection unit 15 is a detection unit that detects the load acting on the material Z when the slider 11 presses the material Z. The load detection unit 15 is a load detection unit that detects the load acting on the material Z, for example, by detecting the current of the servo motor 12. The load detection unit 15 then outputs the detection result to the control device 1 as the detection result of the load acting on the material Z.
[0045] In addition to the above description, the load detection unit 15 may be a load detection unit that includes a strain gauge provided on a punch (not shown) that transmits at least a portion of the load applied by the slider 11 to the material Z, and detects the load acting on the material Z by detecting the amount of strain using the strain gauge.
[0046] Furthermore, the load detection unit 15 detects the load acting on the material Z, thereby making it possible to detect the stress occurring in the material Z during press processing. That is, since it is impossible to measure stress directly, the control device 1 of this embodiment measures the load as a physical property value in place of stress and uses it to control the servo press 10.
[0047] <Configuration of the control device 1> As shown in FIG. 1, the control device 1 includes a predictor 3, a storage unit 4, and a control unit 5.
[0048] The predictor 3 inputs time series data of the load detection results from the load detection unit 15 at each predetermined sampling period, and calculates a load prediction value, which is a predicted value of the load acting on the material Z after a predetermined time, from the input time series data of the load detection results, and outputs it to the control unit 5.
[0049] The predictor 3 has, for example, a neural network constructed as a prediction model (learning model) in which explanatory variables are associated with objective variables, with N pieces (N is an integer of 1 or more) of time series data of load detection results during the first period being the time series data for each sampling period of the load detection results of one press working input from the load detection unit 15, and with M pieces (M is an integer of 1 or more) of load predicted value after a predetermined time being the objective variable. Furthermore, the predictor 3 may continue machine learning based on the constructed learning model, and the learning model may be updated sequentially.
[0050] Specifically, the predictor 3 has a learning model that is generated by performing machine learning using as teacher data a set of data within a first period (i.e., time series data of the above N load detection results) and data a predetermined time after the end of the first period (data of M load detection results), out of the time series data of the load detection results during the stopping operation, and that uses data within the first period as input and outputs data a predetermined time after the end of the first period as the load prediction value.
[0051] Here, a more specific example of the configuration of the predictor 3 will be described with reference to Fig. 4. Fig. 4 is a diagram for explaining a specific example of the configuration of the predictor shown in Fig. 1.
[0052] As shown in FIG. 4, the predictor 3 includes a buffer 3a that sequentially acquires data T(t) of the load detection result from the load detection unit 15, and a learning model 3b that is connected to the buffer 3a and configured using a neural network.
[0053] 4, the buffer 3a holds data T(t-N+1) to T(t) within the first period from the load detection unit 15. The learning model 3b calculates a load prediction value YT(t) for a predetermined time (M units) after the time point t from the data T(t-N+1) to T(t) within the first period input from the buffer 3a, and outputs the load prediction value YT(t) to the control unit 5.
[0054] In addition to the above description, the predictor 3 may be configured using, for example, a recurrent neural network (RNN) as a learning model instead of the buffer 3a and the learning model 3b configured using a neural network.
[0055] The memory unit 4 stores various data used by the control unit 5. The memory unit 4 may also store various software programs that are executed by a computer to cause the computer to function as the control unit 5 or the predictor 3. The memory unit 4 also stores data related to the operation of the servo press 10 that is acquired from the servo press 10 by the control unit 5 and machine-learned. The memory unit 4 also stores in advance data such as a first threshold value, a second threshold value, and a maximum bottom dead center stop time, which are used for a convergence determination, which are input via an operation unit (not shown), as described below.
[0056] The control unit 5 is a computing device having a function of comprehensively controlling each part of the control device 1. The control unit 5 may control each part of the control device 1 by, for example, one or more processors (e.g., CPU, etc.) executing programs stored in one or more memories (e.g., RAM, ROM, etc.).
[0057] 2, the control unit 5 dynamically changes the bottom dead center stop time using the load prediction value from the predictor 3. That is, the control unit 5 is provided with a command value generation function 5a that generates a command value to the servo press machine 10, a position control function 5b that controls the position of the slider 11, a speed control function 5c that controls the rotation speed of the servo motor 12, a torque control function 5d that controls the torque of the servo motor 12, and a convergence determination function 5e that determines the convergence state of the load acting on the material Z.
[0058] In addition, when the control unit 5 determines, using the load detection result from the load detection unit 15, that the load acting on the material Z has converged, it outputs an instruction signal (command value) to the servo press machine 10, instructing the slider 11, which is in a stopped operation, to perform an upward operation.
[0059] Moreover, when the control unit 5 receives a load prediction value from the predictor 3, the control unit 5 uses the load prediction value to determine whether or not the load acting on the material Z is in a converged state. Furthermore, as will be described later in detail, the control unit 5 uses a second threshold value preset in the memory unit 4 to determine whether or not the load acting on the material Z is in a converged state.
[0060] Incidentally, a case in which the load prediction value is not input from the predictor 3, that is, the predictor 3 is not provided in the control device 1, will be described later in a second embodiment.
[0061] §3 Example of operation <Convergence state determination operation> With reference to Fig. 5 as well, an example of the operation of determining the convergence state of the load acting on the material Z in the control device 1 of this embodiment will be specifically described. Fig. 5 is a graph showing a specific example of a change in the load detection result in the servo press machine. Note that the horizontal axis of Fig. 5 indicates time corresponding to a step in the press process, and the vertical axis indicates the load (arbitrary unit).
[0062] For example, when the control device 1 causes a servo press machine 10 having servo motors 12 for four axes to press the same material Z, the load acting on the material Z fluctuates in the servo motors 12 for the four axes as shown by the curves K1, K2, K3, and K4 in Fig. 5. Then, while the slider 11 is stopping, when the deformation of the material Z is completed in each of the four axes, the load shows a substantially constant value in the areas surrounded by the dotted circles KS1, KS2, KS3, and KS4 in Fig. 4.
[0063] In addition, the axis on which the load fluctuation is shown in curve K4 is the axis located closest to the die used for press processing, and the load acting on material Z from that axis converges the slowest, so the load detection result from the load detection unit 15 provided on that axis is used to determine the convergence state in the control device 1.
[0064] In the control device 1 of this embodiment, the control unit 5 determines that the load acting on the material Z has converged when the load detection result from the load detection unit 15 continues to indicate a substantially constant value. Then, the control unit 5 causes the servo press 10 to raise the slider 11, which is currently stopped, to prepare for the press processing of the next material Z.
[0065] Furthermore, the control unit 5 appropriately sets the maximum value of the bottom dead center stop time depending on the material Z. As a result, even if the control unit 5 is unable to appropriately perform the operation of determining the convergence state, it can forcibly end the stopping operation of the slider 11 (bottom dead center stop time) as shown in step S13 in Fig. 10 described later, for example.
[0066] However, if the maximum value of the bottom dead center stop time is reduced, the springback of the material Z will be large, which may result in a decrease in the finishing accuracy of the product P.
[0067] Therefore, in the control device 1 of this embodiment, the maximum value of the bottom dead center stop time (time threshold value) is determined based on, for example, a test operation by a skilled worker. As a result, even when the same material Z is pressed, the control device 1 of this embodiment can eliminate adverse effects due to variations in thickness of each material Z, and ensure the finishing precision of the product P while suppressing an increase in the cycle time of the product P and suppressing a decrease in the productivity of the servo press 10.
[0068] <Prediction and discrimination operations> With reference to Fig. 6 to Fig. 8, a specific example of the predictive operation by the predictor 3 in the control device 1 of this embodiment and the discrimination operation in the control unit 5 using this predictive operation will be described. Fig. 6 is a diagram for explaining an example of the operation of the predictor included in the control device. Fig. 7 is a diagram for explaining an example of the operation of the control unit included in the control device. Fig. 8 is a diagram showing a specific formula used for convergence discrimination in the control unit.
[0069] In Fig. 6, curve K5 shows an example of fluctuation of the load acting on material Z. Also, in Fig. 6, time series data of the load detection results of N press workings (N is an integer equal to or greater than 1), for example, data T(t-N+1) to T(t) within the first period shown in Fig. 4, is input as explanatory variables to learning model 3b of predictor 3. This time series data of the load detection results of N press workings is not limited to the number of times press workings were performed consecutively, and may be determined using importance analysis using Random Forest, for example.
[0070] The learning model 3b of the predictor 3 calculates a load prediction value YT(t) after M press operations (M is an integer equal to or greater than 1), and outputs the calculated load prediction value YT(t) after M operations as a response variable to the control unit 5. In other words, the learning model 3b of the predictor 3 is configured to perform machine learning based on the explanatory variables and output the response variable.
[0071] Next, the control unit 5 performs a convergence judgment as to whether or not the load is in a converged state, using equation (1) shown in Fig. 8. In Fig. 8, YT(tk) is a load prediction value k units before time t, and T(t) is a load detection result at time t. Furthermore, R is a value of the number of samples of data used for the convergence judgment (R is an integer equal to or greater than 1), and ε is a threshold value for the convergence judgment.
[0072] Specifically, the load detection result from the load detection unit 15 fluctuates as shown in curve 70 in FIG. 7, with a value T(t0) at the first appearance point t0 of the convergence value that determines the convergence state (the point at which the load detection result from the load detection unit 15 becomes an approximately constant value), a value T(t) at a point t prior to the first appearance point t0, and a value T(t1) at a point t1 M pieces after the first appearance point t0.
[0073] On the other hand, for the load detection result of curve 70, if the prediction by the predictor 3 is sufficiently accurate at the time point before the first appearance point t0, the load prediction value YT(tM)=T(t) is established. Note that, as shown in curve 71 in Fig. 7, the load prediction value fluctuates like a value YT(t) at time point t, a value YT(t0) at the first appearance point t0, and a value YT(t1) at time point t1.
[0074] As shown in formula (1), the control unit 5 determines that the load acting on the material Z is in a converged state when the difference between the load prediction value newly acquired from the predictor 3 and the average value of the load prediction value within a predetermined period in the past is equal to or less than a second threshold value ε set in advance. For example, as shown by the double-headed arrow in Fig. 7, the control unit 5 can use the load prediction value to predict that the load will converge at an early stage, that is, at time T0 that is M steps before time T1 at which the convergence determination is performed using the load detection result.
[0075] Depending on the setting values of the above M, N, or R, there is a risk of a decrease in the prediction accuracy of the prediction model or the accuracy of judging the convergence state. For this reason, the control device 1 may be configured to calculate an evaluation value (e.g., root mean square error (RMSE)) of the prediction accuracy of the prediction model based on the load detection result and the load prediction value, and reconstruct the prediction model according to the calculation result. In other words, the control device 1 may be configured to perform machine learning taking into account the evaluation value.
[0076] Further, instead of the configuration for determining the convergence state using the above formula (1), for example, a configuration may be used in which a difference value between successive load prediction values is calculated and the convergence state is determined based on this difference value.
[0077] The above M, which is a parameter for the predictor 3 to make a prediction, and the above R, which is a parameter for the control unit 5 to use for the convergence determination, may be different values from each other, or may be the same value.
[0078] <Example of operation of control device 1> An operation example of the control device 1 of this embodiment will be specifically described with reference to Fig. 9 and Fig. 10. Fig. 9 is a flowchart for explaining an operation example of the control device. Fig. 10 is a flowchart for explaining another operation example of the control device.
[0079] 9, first, in step S1, the control device 1 sets a maximum value of the bottom dead center stop time in response to a user's operation. Specifically, before pressing the material Z, for example, a trial operation is performed by a skilled worker, and among the variations in the bottom dead center stop time that change for each pressing due to the thickness of the material Z and the elastic deformation and thermal expansion occurring in the slider 11 of the servo press 10, the bottom dead center stop time that is the maximum time is stored in advance in the memory unit 4 as the maximum value (time threshold value) of the bottom dead center stop time.
[0080] Next, in step S2, the control device 1 causes the servo press machine 10 to perform a normal press operation (press processing) and collects time-series data of the load detection results.
[0081] Next, in step S3, the control device 1 constructs a learning model in the predictor 3 according to the values of N, M, and ε set by the user.
[0082] After constructing the learning model in the predictor 3, the control device 1 causes press working to be performed using the constructed learning model. The press working process using the control device 1 will be described below. In step S11 of FIG. 10, the servo press 10 is caused to perform a downward movement of the slider 11.
[0083] Next, in step S12, the control device 1 causes the servo press machine 10 to perform a stopping operation of the slider 11.
[0084] Next, in step S13, the control device 1 determines whether or not the bottom dead center stop time in the stopping operation has exceeded the maximum value stored in the memory unit 4. If the control device 1 determines that the bottom dead center stop time in the stopping operation has exceeded the maximum value (YES in S13), the control device 1 determines that it is not necessary to continue the stopping operation, and proceeds to step S14. That is, the control device 1 forcibly ends the stopping operation and causes the lifting operation to be performed.
[0085] On the other hand, when the control device 1 determines that the bottom dead center stop time in the stopping operation does not exceed the maximum value (NO in S13), the predictor 3 calculates a load prediction value based on the prediction model (step S15).
[0086] Next, in step S16, the control unit 5 determines whether or not the load detection result from the load detection unit 15 is in a converged state. If the control unit 5 determines that the load detection result is not in a converged state (NO in S16), the control device 1 determines that it is necessary to continue the stopping operation, and proceeds to step S12.
[0087] On the other hand, when the control section 5 determines that the load detection result has converged (YES in S16), the control device 1 determines that it is not necessary to continue the stopping operation, and proceeds to step S14.
[0088] Then, in step S14, the control device 1 causes the servo press 10 to end the stopping operation of the slider 11 and perform the raising operation.
[0089] As described above, the control device 1 of this embodiment uses the load detection result from the load detection unit 15 during the stopping operation of the slider 11 to determine whether or not the load acting on the material Z has converged. Then, when the control device 1 determines that the load acting on the material Z has converged, it ends the stopping operation of the slider 11 and immediately causes the slider 11 to ascend. As a result, the control device 1 of this embodiment can appropriately change the bottom dead center stop time in the above-mentioned stopping operation, and can dynamically change the bottom dead center stop time of the slider 11 for the servo press 10 according to the load acting on the material Z.
[0090] Here, the effect of the control device 1 of this embodiment will be specifically described with reference to Fig. 11. Fig. 11 is an explanatory diagram for explaining a specific example of the effect of the control device. Note that the horizontal axis of Fig. 11 indicates the time corresponding to the step in the press working, and the vertical axis indicates the load (arbitrary unit) and the position of the slider 11 (arbitrary unit).
[0091] 11, when the servo press 10 performs press processing, the load detection result from the load detection unit 15 fluctuates as shown by the dashed line K6. Here, when the control unit 5 determines that the load acting on the material Z is in a converged state at time T10, the control device 1 causes the servo press 10 to end the stopping operation of the slider 11 at time T11 and start the lifting operation.
[0092] As a result, in the servo press 10, the slider 11 rises from the bottom dead center position to the top dead center position at time T11, as shown by the curve S1. Then, in the servo press 10, the slider 11 becomes capable of performing press processing on the next material Z at time T13.
[0093] On the other hand, in a comparative example in which the bottom dead center stop time is not dynamically changed, for example, when the bottom dead center stop time ends at time T12, in the servo press machine of the comparative example, the slider rises from the bottom dead center position to the top dead center position at time T12, as shown by curve S2. Therefore, in the comparative example, the slider becomes able to perform press processing on the next material Z at time T14.
[0094] That is, as shown in FIG. 11, the control device 1 of this embodiment can shorten the cycle time by the difference between time T14 and time T13 while maintaining the finishing precision of the product P, as compared to the comparative example.
[0095] Specifically, the control device 1 of this embodiment determines that the load acting on the material Z has converged, and therefore plastic deformation is complete in the pressed material Z. In other words, by making the above determination, the control device 1 of this embodiment detects that the material Z is in a state in which springback is unlikely to occur, and therefore the finishing precision of the product P can be easily ensured.
[0096] Furthermore, in the control device 1 of this embodiment, even if there is variation in the material Z, for example, in the case where the material Z is a plate-like member and there is variation in its thickness, it is detected that the springback is unlikely to occur in each material Z as described above. Therefore, in the control device 1 of this embodiment, even if there is variation in the material Z, it is possible to easily ensure the finishing precision of the product P.
[0097] Furthermore, in the control device 1 of this embodiment, since machine learning by the predictor 3 is used, it is possible to determine earlier that the load of the material Z is in a converged state, and it is possible to perform convergence determination at an early stage. As a result, in the control device 1 of this embodiment, it is possible to easily shorten the cycle time.
[0098] [Embodiment 2] Another embodiment of the present disclosure will be described below with reference to Fig. 12. Fig. 12 is a block diagram showing a configuration example of a control device and a servo press machine according to a second embodiment of the present disclosure. For convenience of explanation, the same reference numerals are given to members having the same functions as those described in the above embodiment, and the explanations thereof will not be repeated.
[0099] The main difference between the first and second embodiments is that the predictor 3 is omitted in the control device 1.
[0100] In the control device 1 of the second embodiment, the convergence determination function 5e of the control unit 5 receives the load detection result from the load detection unit 15 instead of the load prediction value YT(t) output by the predictor 3 in the first embodiment.
[0101] Then, the control unit 5 of the control device 1 of the second embodiment determines that the load acting on the material Z is in a converged state, for example, when the difference between the load detection result newly acquired from the load detection unit 15 and the average value of the load detection results within a predetermined period of time is less than a preset first threshold value. As a result, in the control device 1 of the second embodiment, similarly to the first embodiment, the bottom dead center stop time in the above-mentioned stopping operation can be appropriately changed, and the bottom dead center stop time of the slider 11 can be dynamically changed for the servo press 10 in accordance with the load acting on the material Z.
[0102] As a result, in the control device 1 of the second embodiment, similarly to the first embodiment, it is possible to ensure the finishing precision of the product P in the servo press machine 10 while shortening the cycle time.
[0103] [Software implementation example] The functional blocks of the control device 1 (particularly, the control unit 5) may be realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like, or may be realized by software.
[0104] In the latter case, the control unit 5 includes a computer that executes instructions of a program, which is software that realizes each function. The computer includes, for example, one or more processors, and a computer-readable recording medium that stores the program. In the computer, the processor reads the program from the recording medium and executes it, thereby achieving the object of the present disclosure.
[0105] The processor may be, for example, a CPU (Central Processing Unit). The recording medium may be a "non-transitory tangible medium" such as a ROM (Read Only Memory), a magnetic disk, a card, a semiconductor memory, a programmable logic circuit, or the like. The device may further include a RAM (Random Access Memory) for expanding the program.
[0106] The program may be supplied to the computer via any transmission medium capable of transmitting the program (such as a communication network or a broadcast wave). Note that one aspect of the present invention may also be realized in the form of a data signal embedded in a carrier wave, in which the program is embodied by electronic transmission.
[0107] The present disclosure is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present disclosure. [Explanation of symbols]
[0108] 1 Control device 3. Predictor 5. Control section 10. Servo press machine 11 Slider 12 Servo motor 15 Load detection unit
Claims
1. A control device for controlling a servo press machine which performs press working on a material by moving a slider in a vertical direction, the control device comprising: a servo motor which drives the slider; a position detection unit which detects a position of the slider; and a load detection unit which detects a load acting on the material, a control unit that controls the servo motor using a position detection result by the position detection unit and a load detection result by the load detection unit, The control unit is a descending operation of descending the slider toward a bottom dead center position, which is a lowest point position of the slider; a stopping operation for stopping the slider at the bottom dead center position; a lifting operation of lifting the slider from the bottom dead center position as a series of steps in the press working; During the stopping operation, based on the load detection result, it is determined whether or not the load acting on the material has converged; a control device that performs the lifting operation when it is determined that the load acting on the material has converged.
2. The control device according to claim 1, wherein the control unit determines that the load acting on the material is in a converged state when a difference between the load detection result newly acquired from the load detection unit and an average value of the load detection result within a predetermined period of time is less than a predetermined first threshold value.
3. A predictor that obtains the load detection result from the load detection unit, calculates a load prediction value that is a prediction value of the load acting on the material after a predetermined time from time series data of the load detection result obtained, and outputs the load prediction value to the control unit, The control unit is During the stopping operation, The control system of claim 1 , wherein the load prediction value is used to determine whether the load acting on the material is in a converged state.
4. 4. The control device according to claim 3, wherein the control unit determines that the load acting on the material is in a converged state when a difference between the load prediction value newly acquired from the predictor and an average value of the load prediction value within a predetermined period of time in the past is less than a preset second threshold value.
5. The predictor may further comprise: The control device according to claim 3 or 4, further comprising a learning model generated by performing machine learning using a pair of data within a first period and data from the end of the first period as teacher data, among the time series data of the load detection result during the stopping operation, as teacher data, to generate a learning model that uses data within the first period as input and outputs data after the specified time from the end of the first period as the load prediction value.
6. The control device according to claim 1 , wherein the load detection unit detects a torque of the servo motor to thereby detect a load acting on the material.
7. 6. The control device according to claim 1, wherein the load detection unit includes a strain gauge provided on a punch that transmits at least a portion of the load applied by the slider to the material, and detects the load acting on the material by detecting an amount of strain using the strain gauge.
8. A control method for controlling a servo press machine that performs press processing on a material by moving a slider in a vertical direction, the servo press machine including a servo motor that drives the slider, a position detection unit that detects a position of the slider, and a load detection unit that detects a load acting on the material, the method comprising: a lowering operation step of lowering the slider toward a bottom dead center position, which is a lowest point position of the slider; a stopping operation step of stopping the slider at the bottom dead center position; a raising operation step of raising the slider from the bottom dead center position, During the stopping operation step, a substep of determining whether or not the load acting on the material is in a converged state based on a load detection result by the load detection unit is repeatedly executed; The control method includes transitioning to the lifting operation step when it is determined that the load acting on the material has converged.
9. 2. A control program for causing a computer to function as the control device according to claim 1, the control program causing a computer to function as the control unit.
Citation Information
Patent Citations
Brick press molding method
JP1990074304A
Quality controlling method of press machine and its device
JP1995164199A
Electric cylinder control method and electric cylinder control system
JP2011098350A
Electrically-driven press working machine
JP2016120522A
