JP6961130B1 - Simulation program, simulation device, and simulation method - Google Patents

Abstract

A machine model that simulates the operation of a machine and a driver model that simulates the operation of a machine, a machine equipped with a motor, a motor driver that controls the motor, and a controller that controls the motor driver. This is a simulation program that simulates using Let the computer perform a model selection step to select the machine model and driver model to be used, and a model calculation step to simulate the operation of the machine system using the selected machine model and driver model, and make one machine a different abstraction. Multiple machine models simulated by degree are included in the machine model candidates, or a plurality of driver models simulating one motor driver with different degrees of abstraction are included in the driver model candidates.

Description

The present disclosure relates to a simulation program, a simulation apparatus, and a simulation method for simulating the operation of at least one of the operation of a machine and the operation of a motor driver.

In order for a machine driven by a motor to perform a desired operation, it is necessary to adjust parameters such as design parameters that determine the configuration of the machine and control parameters for controlling the motor. Repeating the trial production and test operation of the machine for adjusting these parameters requires enormous labor, time, and cost. Therefore, the parameters are adjusted by using the calculation result of the numerical calculation by the simulation device that simulates the operation of the machine.

In order to simulate the operation of a machine, a model for simulating the operation of the machine is required on software. Furthermore, there are a wide variety of machines with different specifications for each application, and it is necessary to create a model of the machine according to the specifications. For example, a model used for a simulation for adjusting parameters is required to have accuracy, but a simulation using a model that simulates the entire machine with high accuracy requires a huge calculation cost. Therefore, a model with an appropriate degree of abstraction is used for each part, such as simulating the necessary part with high accuracy according to the purpose of the simulation and simplifying the other parts.

In the simulator described in Patent Document 1, models such as a motor driver for controlling a motor and a sensor signal unit for inputting a sensor signal are created, and by combining these models, models can be used for various machines. It is easy to create.

International Publication No. 2008/120304

However, in the technique of Patent Document 1, although it is possible to create a model for machines having different configurations, it is not possible to use a model having a different degree of abstraction for one machine. Therefore, in the case of the technique of Patent Document 1, the user needs to create a model with an appropriate degree of abstraction every time the content to be simulated changes, and it takes time and effort to create the model, so that the calculation efficiency is good. There was a problem that the simulation could not be realized easily.

The present disclosure has been made in view of the above, and an object of the present disclosure is to obtain a simulation program capable of easily realizing a simulation with high calculation efficiency.

In order to solve the above-mentioned problems and achieve the object, the present disclosure discloses a mechanical system including a machine provided with a motor and driven by the motor, a motor driver for controlling the motor, and a controller for controlling the motor driver. Is a simulation program that simulates the operation of From the library, the machine model used when simulating the operation of the machine and the operation of the motor driver are simulated based on the degree of abstraction setting, which is the information in which the degree of abstraction of the simulation when simulating the operation of the machine system is set. Have the computer perform a model selection step to select the driver model to be used. In addition, the simulation program causes the computer to execute a model calculation step that simulates the operation of the mechanical system using the machine model and the driver model selected in the model selection step. If at least one of the machine model candidates and the driver model candidates is plural, and there are multiple machine model candidates, the machine model candidates include a plurality of machine models that simulate one machine with different abstractions. Is included, and when there are a plurality of driver model candidates, the driver model candidates include a plurality of driver models that simulate one motor driver with different abstractions.

The simulation program according to the present disclosure has an effect that a simulation with high calculation efficiency can be easily realized.

The figure which shows the structure of the mechanical system whose operation is simulated by the simulation apparatus which concerns on Embodiment 1. The figure which shows the structure of the simulation apparatus which concerns on Embodiment 1. A flowchart showing a simulation processing procedure by the simulation apparatus according to the first embodiment. The figure which shows the structure of the simulation apparatus which concerns on Embodiment 2. The figure for demonstrating the sequence program used in the mechanical system which simulates the operation by the simulation apparatus which concerns on Embodiment 2. The figure which shows the structure of the simulation apparatus which concerns on Embodiment 3. The figure for demonstrating the neural network used by the simulation apparatus which concerns on Embodiment 3. The figure which shows the structure of the simulation apparatus which concerns on Embodiment 4. The figure for demonstrating the process which the simulation apparatus which concerns on Embodiment 4 generates a new model by simplification of a model. The figure which shows the structure of the simulation apparatus which concerns on Embodiment 5. The figure for demonstrating the neural network used by the simulation apparatus which concerns on Embodiment 5. The figure which shows the structure of the mechanical system whose operation is simulated by the simulation apparatus which concerns on Embodiment 6. The figure which shows the structure of the simulation apparatus which concerns on Embodiment 6. The figure which shows the hardware configuration example which realizes the simulation apparatus which concerns on Embodiment 1.

Hereinafter, the simulation program, the simulation apparatus, and the simulation method according to the embodiment of the present disclosure will be described in detail with reference to the drawings.

Embodiment 1.
FIG. 1 is a diagram showing a configuration of a mechanical system whose operation is simulated by the simulation apparatus according to the first embodiment. In the first embodiment, a case where the simulation device executes a simulation which is a process of simulating the operation of the mechanical system 20A having the three-axis motor shown in FIG. 1 will be described.

The mechanical system 20A includes a machine 1 driven by a motor and a controller 10A that controls the machine 1. Further, the machine 1 has a stage 5 and a head 6. In the following description, the two directions in the plane parallel to the upper surface of the stage 5 and orthogonal to each other are defined as the X direction and the Y direction. Further, the direction orthogonal to the X direction and the Y direction is defined as the Z direction.

The machine 1 has a three-axis motor including an X-axis 1a extending in the X direction, a Y-axis 1b extending in the Y direction, and a Z-axis 1c extending in the Z direction.

The stage 5 is connected to the X-axis 1a and the Y-axis 1b, and moves in the XY plane by the operation of the X-axis 1a and the operation of the Y-axis 1b. The head 6 is connected to the Z-axis 1c and moves in the Z-axis 1c direction by the operation of the Z-axis 1c.

The controller 10A controls the positions of the stage 5 in the X direction and the Y direction by controlling the operation of the X axis 1a and the operation of the Y axis 1b. Further, the controller 10A controls the position of the head 6 attached to the tip of the Z-axis 1c in the Z direction by controlling the operation of the Z-axis 1c. The machine 1 processes the work Wp placed on the stage 5 by the head 6.

Further, the mechanical system 20A is equipped with a motor driver (not shown). In the mechanical system 20A, the controller 10A controls the motor driver, and the motor driver drives and controls the motor of each axis. Examples of the controller 10A are a programmable logic controller, an industrial personal computer, a servo system controller and the like.

FIG. 2 is a diagram showing a configuration of a simulation device according to the first embodiment. The simulation device 11 is a computer that simulates the operation of the mechanical system 20A.

The simulation device 11 includes a model selection unit 12, a model library 13, and a model calculation unit 14. The model selection unit 12 selects a model corresponding to the abstraction degree setting from the model candidates stored in the model library 13 based on the abstraction degree setting given from the outside.

The abstraction level setting is information indicating the abstraction level of the simulation. In other words, the abstraction degree setting is information for selecting one from the simulation options prepared in advance. That is, the abstraction level setting contains information that determines the abstraction level of the model. The abstraction degree setting is input to the model selection unit 12 by the user of the simulation device 11.

The abstraction degree setting may be information indicating the purpose of the simulation, or may be information indicating the abstraction degree numerically. When the abstraction degree setting is information indicating the abstraction degree numerically, the larger the abstraction degree numerical value is, the higher the abstraction degree of the simulation corresponds to the simplified simulation. The abstraction level setting may be information indicating the purpose of the simulation.

The model selection unit 12 stores the correspondence information indicating the correspondence between the abstraction degree setting and the selected model. The model selected by the model selection unit 12 is a model corresponding to the purpose of the simulation. Therefore, it can be said that the correspondence information is the information in which the abstraction degree setting and the model corresponding to the purpose of the simulation are associated with each other. The model selection unit 12 selects a model corresponding to the abstraction degree setting based on the correspondence information.

The model library 13 stores a plurality of driver models having different abstractions. In the first embodiment, the case where the driver models stored in the model library 13 are the driver models D1 and D2 will be described.

The driver models D1 and D2 are models that simulate the operation of the motor driver. The driver model D1 is a model that simulates the operation of the motor driver with higher accuracy than the driver model D2. The driver model D1 simulates the functions of, for example, a feedforward controller, a position feedback controller, a speed feedback controller, and a current controller. Further, the driver model D2 is a model that simulates the operation of the motor driver more simply than the driver model D1. The driver model D2 simulates only the function of the feedforward controller, for example.

Further, the model library 13 stores a plurality of machine models having different abstractions. In the first embodiment, a case where the machine model stored in the model library 13 is two machine models M1 and M2 will be described.

The machine models M1 and M2 are models that simulate the operation of the machine 1. The machine model M1 is a model that simulates the operation of the machine 1 with higher accuracy than the machine model M2. The machine model M1 simulates, for example, the vibration characteristics of the machine 1. Further, the machine model M2 is a model that simulates the operation of the machine 1 more simply than the machine model M1. The mechanical model M2 simulates, for example, the operating range of the stage 5 and the head 6. Details of the driver models D1 and D2 and the mechanical models M1 and M2 will be described later.

The model selection unit 12 may select at least one of the driver model and the mechanical model. For example, the model selection unit 12 may select only the driver model or only the mechanical model. In other words, at least one of the machine model candidate and the driver model candidate held by the model library 13 is plural. When there are a plurality of machine model candidates, the machine model candidates include a plurality of machine models that simulate one machine with different abstractions. When there are a plurality of driver model candidates, the driver model candidates include a plurality of driver models that simulate one motor driver with different abstractions.

When selecting only the driver model from the plurality of driver models, the model selection unit 12 selects one fixed machine model. On the other hand, when the model selection unit 12 selects only a machine model from a plurality of machine models, the model selection unit 12 selects one fixed driver model.

Further, the model selection unit 12 may select a combination of the driver model and the machine model. In this case, the model selection unit 12 stores the correspondence information indicating the correspondence between the abstraction degree setting and the combination of the selected models. In the first embodiment, a case where the model selection unit 12 selects a combination of the driver model and the machine model will be described. In the following description, the combination of the driver model and the machine model may be referred to as a model set. The model selection unit 12 notifies the model calculation unit 14 of the model set selected from the model library 13.

The model set corresponds to the purpose of the simulation. In the first embodiment, a case where two types of options are prepared as a simulation will be described. For example, when the simulation device 11 wants to execute a simulation for confirming the movable range of the movable portion of the mechanical system 20A, a model of a motor for driving the movable portion and a model of a motor driver for driving the motor is required. In this case, the simulation device 11 does not need to simulate the delicate vibration characteristics and the control response characteristics of the motor with high accuracy, and the motor model, the driver model, and the mechanical model may be abstracted.

On the other hand, in order for the user to adjust the control parameters of the motor driver of the multi-axis device equipped with a large number of motors, a precise model of the motor driver, the motor, and the machine 1 to be adjusted is required. In this case, the simulation device 11 does not need a precise model for the motor driver and the motor that are not the adjustment targets.

Further, a simulation for debugging the sequence program of the controller 10A included in the mechanical system 20A such as a programmable logic controller, an industrial personal computer, and a servo system controller may be performed. In this case, the simulation device 11 requires a model that accurately simulates the execution cycle and execution order of the sequence program, but the vibration characteristics of the machine 1 may be simplified. In addition, there are various purposes of simulation, such as simulation for thermal analysis of mechanical system 20A driven by a motor and simulation of power consumption analysis. It is desirable to use it to perform a simple simulation.

Therefore, the simulation device 11 stores various models corresponding to various simulation purposes, and executes the simulation using the models corresponding to the selected simulation purposes.

In the first embodiment, for example, there are two types of simulation options: a simulation for confirming the operating range of the stage 5 and the head 6, and a simulation for adjusting the parameters of the motor driver that moves the stage 5. The case where it is prepared will be described.

The model selection unit 12 selects a model suitable for the purpose of the simulation from the model library 13 based on the abstraction degree setting. Here, the plurality of machine models M1 and M2 held by the model library 13 are software having a common interface. That is, the machine models M1 and M2 are software in which input / output is common information.

The user may define, for example, an object that collectively stores motor information such as the motor position and the motor speed, and use this object as an input for a common interface of the machine model. Further, the user may define an object that summarizes information such as the position of the stage 5, the position of the head 6, the external force applied to the motor, and the heat generated by the machine 1, and use this object as the output of the common interface of the machine model. .. In this case, the type of input / output may be a variable type or a structure type. The common interface of the machine model may be common among different machine models M1 and M2 that simulate one machine 1. The input / output specifications for the common interface of the machine model are not limited to the contents described in the first embodiment.

In the first embodiment, the common interface of the mechanical models M1 and M2 has an input interface capable of acquiring the motor position of the X-axis 1a, the motor position of the Y-axis 1b, and the motor position of the Z-axis 1c. It shall be.

The machine model M1 is, for example, a model that considers the dynamics of the machine 1 and simulates the vibration characteristics of the machine 1. For example, the machine model M2 changes the position of the stage 5 and the position of the head 6 only by kinematics based on the motor position without considering the dynamics of the machine 1, and simulates the operating range of the stage 5 and the head 6. The mechanical model M2 may simulate only the operating range of one of the stage 5 and the head 6.

The model calculation unit 14 reads the driver model and the machine model corresponding to the model set notified from the model selection unit 12 from the model library 13. The model calculation unit 14 performs a simulation calculation of the motor driver and the machine 1 by using the read driver model and the machine model.

The model calculation unit 14 calculates the model of the mechanical system 20A including the model selected from the model library 13. For example, assume that the model of the mechanical system 20A includes a motor model and a controller model. The motor model is a model that simulates the operation of the motor, and the controller model is a model that simulates the operation of the controller 10A. In this case, the model calculation unit 14 includes a driver model and a motor model corresponding to the X-axis 1a, a driver model and a motor model corresponding to the Y-axis 1b, a driver model and a motor model corresponding to the Z-axis 1c, and a mechanical model. The model of the mechanical system 20A includes a controller model that gives a position command of the motor to the motor driver. The model calculation unit 14 may acquire the motor model and the controller model from the model library 13, or may acquire from other than the model library 13. The model calculation unit 14 outputs the simulation result to an external device such as a display device. As a result, the display device displays the simulation result.

Next, the simulation processing procedure by the simulation device 11 will be described. FIG. 3 is a flowchart showing a simulation processing procedure by the simulation apparatus according to the first embodiment.

The model selection unit 12 of the simulation device 11 accepts the abstraction degree setting input by the user (step S10). The model selection unit 12 selects a model set from the model library 13 based on the abstraction degree setting (step S20). The model selection unit 12 selects, for example, a model set including two models, one of the driver models D1 and D2 and one of the mechanical models M1 and M2.

The model selection unit 12 notifies the model calculation unit 14 of the selected model set. The model calculation unit 14 reads the model set selected by the model selection unit 12 from the model library 13 (step S30).

The model calculation unit 14 simulates the operation of the mechanical system 20A using the read model set (step S40). The model calculation unit 14 outputs the simulation result to a display device or the like (step S50). This allows the user to check the simulation result.

Next, the detailed operation of the simulation device 11 according to the first embodiment will be described. First, consider a case where the user selects a simulation for the purpose of confirming the operating range of the stage 5 and the head 6 as the abstraction degree setting.

When simulation for the purpose of confirming the operating range of the stage 5 and the head 6 is selected as the correspondence information of the model selection unit 12, 3 corresponding to the X-axis 1a, the Y-axis 1b, and the Z-axis 1c. It is assumed that it is predetermined to select one driver model D2 and a mechanical model M2. In this case, when the user is specified the abstraction degree setting corresponding to the simulation for the purpose of confirming the operating range of the stage 5 and the head 6, the model selection unit 12 is based on the abstraction degree setting and the correspondence information. The driver model D2 is selected as the driver model corresponding to the X-axis 1a, the Y-axis 1b, and the Z-axis 1c. Further, the model selection unit 12 selects the machine model M2 as the machine model.

The model calculation unit 14 uses these selected model sets, that is, the driver model D2 and the mechanical model M2, to execute a simulation that simulates the operation of the head position, which is the position of the head 6. At this time, the model calculation unit 14 inputs the position command of the X-axis 1a generated by the controller model to the driver model D2 of the X-axis 1a. In this case, the driver model D2 simulates and controls the motor position of the X-axis 1a only by feedforward control. The model calculation unit 14 inputs the position command generated by the controller model to the driver model D2 for the Y-axis 1b and the Z-axis 1c as well as the X-axis 1a. In this case, the driver model D2 simulates and controls the motor positions of the Y-axis 1b and the Z-axis 1c only by feedforward control. As a result, the model calculation unit 14 calculates the motor positions of the X-axis 1a, the Y-axis 1b, and the Z-axis 1c using the driver model D2.

The model calculation unit 14 calculates the machine model M2 by inputting the motor positions of the X-axis 1a, the Y-axis 1b, and the Z-axis 1c calculated in this way. Here, since the mechanical model M2 is selected as the mechanical model, the stage position and the head position are calculated only by kinematics from the motor positions of each axis. As a result, the model calculation unit 14 can simulate the operating ranges of the stage 5 and the head 6. The model calculation unit 14 outputs the operating ranges of the stage 5 and the head 6 as a result of the simulation.

Next, consider a case where the user selects a simulation for adjusting the parameters of the motor driver of the stage 5 as the abstraction degree setting.

When a simulation for adjusting the parameters of the motor driver of the stage 5 is selected for the correspondence information of the model selection unit 12, the driver model D1 is selected as the driver model corresponding to the X-axis 1a and the Y-axis 1b. It is assumed that it is predetermined to do. When a simulation for adjusting the parameters of the motor driver of the stage 5 is selected for the correspondence information, the driver model D2 is selected as the driver model corresponding to the Z-axis 1c, and the mechanical model is used as the mechanical model. It is assumed that the selection of M1 is predetermined. In this case, when a simulation for adjusting the parameters of the motor driver of the stage 5 is specified to the user, the model selection unit 12 sets the X-axis 1a and the Y-axis 1b based on the abstraction degree setting and the correspondence information. Select driver model D1 as the corresponding driver model. Further, the model selection unit 12 selects the driver model D2 as the driver model corresponding to the Z-axis 1c. Further, the model selection unit 12 selects the machine model M1 as the machine model.

The model calculation unit 14 executes a simulation that simulates the operation of the head position using these selected models, that is, the driver models D1 and D2 and the mechanical model M1. At this time, the model calculation unit 14 inputs the position command of the X-axis 1a generated by the controller model to the driver model D1 of the X-axis 1a. In this case, the driver model D1 simulates and controls the motor position by the feedforward controller, the position feedback controller, the speed feedback controller, and the current controller. The machine 1 is equipped with, for example, a PI (Proportional-Integral) controller as a position feedback controller and a speed feedback controller. In this case, the respective proportional gain and integral gain are control parameters, and the position control performance of the motor of the X-axis 1a is changed by changing these control parameters.

Since the driver model of the Y-axis 1b is the driver model D1 as in the case of the X-axis 1a, the position control performance of the motor of the Y-axis 1b is changed based on the control parameters. Further, since the motor driver model of the Z-axis 1c is the driver model D2, the motor position is controlled only by feedforward control.

Further, as the mechanical model, the mechanical model M1 that simulates the vibration characteristics is selected. Therefore, the operation of the motor causes vibration or the like in the machine 1. Therefore, the model calculation unit 14 outputs the stage position as a simulation result. As a result, the user can confirm the presence or absence of vibration of the stage 5. Further, the user can adjust the control parameters of the X-axis 1a and the Y-axis 1b so that the vibration of the stage 5 does not occur while confirming the presence or absence of the vibration of the machine 1 which is the simulation result.

As described above, in the first embodiment, the user selects one of the purposes of a plurality of types of simulations, for example, two types, and inputs the purpose to the simulation device 11 as an abstraction degree setting. Based on the abstraction degree setting and the correspondence information, the model selection unit 12 makes the required calculation amount smaller than the predetermined amount while having the accuracy required for the simulation from the models stored in the model library 13. A model with an appropriate level of abstraction is selected.

For example, when a simulation for confirming the operating range of the head 6 and the stage 5 is selected, the model selection unit 12 selects the mechanical model M2 and the driver model D2. The mechanical model M2 calculates the stage position and head position corresponding to the motor position based solely on kinematics. Since this mechanical model M2 does not include the influence of dynamics such as an external force on the motor position, there is no need to simulate controlling the motor position by feedback control. Therefore, when selecting the mechanical model M2, the model selection unit 12 selects the simple driver model D2 having only feedforward control.

As a result, the simulation device 11 does not need to calculate a complicated differential equation related to the dynamics of the machine 1 or the feedback control, and the calculation of the model calculation unit 14 becomes very simple. Therefore, the simulation device 11 can confirm the operating range of the head 6 and the stage 5 at a very small calculation cost as compared with the case of calculating the differential equation when executing the simulation of the precise model. That is, the simulation device 11 can execute a simulation using a model having an appropriate degree of abstraction to satisfy the purpose of the simulation while keeping the calculation cost small.

On the other hand, when a simulation for adjusting the parameters of the motor driver of the stage 5 is selected, the model selection unit 12 executes a detailed simulation as a driver model corresponding to the X-axis 1a and the Y-axis 1b that drive the stage 5. Select the driver model D1 to be used. Further, the model selection unit 12 selects the machine model M1 that executes a precise simulation as the machine model. The model selection unit 12 selects the driver model D2, which is a simple driver model, for the driver model of the Z-axis 1c, which does not affect the control performance of the motor drivers of the X-axis 1a and the Y-axis 1b.

Then, the user adjusts the parameters of the X-axis 1a and the Y-axis 1b while checking the vibration of the stage 5 output as the simulation result. In this case, since the simulation device 11 uses a precise model for the operation of the X-axis 1a, the Y-axis 1b, and the stage 5 for the simulation, the user can accurately adjust the parameters. The elaborate model in this case is a model with a large amount of calculation, but the simulation device 11 does not need to consider the adjustment of the control parameters of the motor drivers of the X-axis 1a and the Y-axis 1b. Uses a simplified model. Therefore, the simulation device 11 can reduce the amount of calculation for simulating the operation of the Z-axis 1c. That is, the simulation device 11 can execute a simulation using a model having an appropriate degree of abstraction to satisfy the purpose of the simulation while keeping the calculation cost small.

In this way, regardless of which abstraction degree setting is specified by the user, the simulation device 11 can select an appropriate abstraction degree model according to the purpose of the simulation based on the abstraction degree setting. , Simulation with good calculation efficiency can be easily realized.

For example, when the parameters of the machine 1 driven by the motor are adjusted, the simulation is executed many times while the parameters are changed, so that it is required to keep the calculation cost and the time cost of the simulation small. The simulation device 11 of the first embodiment selects an appropriate abstraction degree model according to the purpose of the simulation based on the abstraction degree setting even when the parameters are changed and the simulation is executed many times. Therefore, the calculation cost and the time cost can be kept small.

Here, the effect that the mechanical models M1 and M2 in the first embodiment have a common interface will be described. In the middle of adjusting the parameters of the motor driver, for example, no matter how the parameters of the motor driver are adjusted, the desired operating specifications of the mechanical system 20A may not be satisfied, and it may be necessary to change the structure, material, etc. of the machine 1. be. In addition, for example, it was newly found that the operating characteristics of the machine 1 change under the influence of the temperature or humidity of the installation environment of the machine 1, and the required simulation accuracy is achieved unless the new machine model incorporates these influences. It may not be possible.

In such cases, the user will create a new machine model. At that time, if a new machine model conforming to the specifications of the common interface of the machine model is created, the simulation of the new machine system 20A can be executed only by exchanging the machine models in the model calculation unit 14. That is, when the machine models are replaced, it is not necessary to add or change a program that executes operations between the models. As described above, since the mechanical models have a common interface, the simulation device 11 can easily execute an appropriate simulation according to the purpose of the simulation.

In the first embodiment, for the sake of simplicity, the case where one of the two types of simulation objectives is selected as the abstraction degree setting has been described, but even if any of the three or more types of simulation objectives is selected, good.

Further, in addition to the purpose of the simulation described in the description of the first embodiment, the simulation device 11 is a driver model for the Z-axis 1c motor driver for the purpose of adjusting the parameters of the Z-axis 1c motor driver, for example. D1 may be selected. In this case, the simulation device 11 may select the driver model D2 as the motor driver for the X-axis 1a and the Y-axis 1b, and select the machine model M1 as the machine model. As a result, the user can efficiently adjust the control parameters of the motor driver of the Z-axis 1c.

Further, in the first embodiment, the case where the control parameter of the motor driver is the gain of PI control has been described, but the control parameter of the motor driver is not limited to the gain of PI control. The simulation device 11 may add a notch filter or a low-pass filter to the feedback controller and use these cutoff frequencies as control parameters. Further, the simulation device 11 may use the response band of the feedforward controller as a control parameter.

Although only one type of motor model is used in the first embodiment, a plurality of models having different abstractions for the motor model may be stored in the model library 13. In this case, the model selection unit 12 selects the motor model from the model library 13 based on the abstraction degree setting and the correspondence information. For example, a high-precision model of a motor model includes a stator model and a rotor model, and simulates rotor eccentricity, cogging, and the like. On the other hand, the simple motor model may be a model in which the rotor operates according to the position command of the motor driver.

Further, as for the driver model, the simulation device 11 may store a driver model having a higher accuracy than the driver model D1 in the model library 13. This highly accurate driver model may be, for example, a driver model that simulates a PWM (Pulse Width Modulation) controller, an inverter switching circuit, or the like.

Further, the simulation device 11 may store a driver model simpler than the driver model D2 in the model library 13. This simple driver model may be a driver model having a wasted time element that delays the position command received from the controller 10A, or may be a driver model that operates only the low-pass filter.

Further, in the first embodiment, the case where the driver model performs position control has been described, but the driver model may be a driver model that performs speed control, torque control, and the like.

Further, among the simulation devices 11, the model selection unit 12, the model calculation unit 14, and the model library 13 do not necessarily have to be mounted on one piece of hardware. For example, among the components of the simulation device 11, the model library 13 may be implemented in the cloud server.

Further, the simulation device 11 may not simulate the entire operation of the machine system 20A, but may simulate the operation of the machine 1. That is, the simulation device 11 does not have to simulate the operation of the controller 10A.

As described above, in the first embodiment, the simulation device 11 operates the machine 1 from the model library 13 that holds a plurality of machine model candidates and a plurality of driver model candidates based on the abstraction degree setting. The operation of the mechanical system 20A is simulated by selecting the mechanical model used for simulating and the driver model used for simulating the operation of the motor driver. As a result, regardless of which abstraction degree setting is specified by the user, the simulation device 11 can select an appropriate abstraction degree model according to the purpose of the simulation based on the abstraction degree setting. Therefore, the simulation device 11 can easily realize a simulation with high calculation efficiency.

Further, since the machine models M1 and M2 have a common interface, when the machine models M1 and M2 are changed, the user does not need to add or change a program that executes an operation between the models. Further, if the machine model created by the user is also created using the common interface, it can be registered in the model library 13 and used.

Embodiment 2.
Next, the second embodiment will be described with reference to FIGS. 4 and 5. In the simulation apparatus of the second embodiment, in addition to selecting the driver model and the mechanical model, the controller model according to the purpose of the simulation is selected from a plurality of controller models having different degrees of abstraction.

FIG. 4 is a diagram showing a configuration of a simulation device according to a second embodiment. Of the components of FIG. 4, components that achieve the same functions as the simulation device 11 of the first embodiment shown in FIG. 2 are designated by the same reference numerals, and redundant description will be omitted.

The mechanical system handled as the target of the simulation in the second embodiment is the same mechanical system 20A as in the first embodiment. That is, the simulation device 21 simulates the operation of the mechanical system 20A in the same manner as the simulation device 11. Also in the second embodiment, one of the two simulation purposes is specified by the abstraction degree setting specified by the user.

One of the simulations selected by the simulation device 21 aims to adjust the control parameters of the motor driver that drives the stage 5 of the mechanical system 20A as in the first embodiment. The other of the selected simulations is intended to confirm the operation of the sequence program that determines the operation order of the stage 5 and the head 6.

The simulation device 21 includes a model selection unit 22, a model library 23, and a model calculation unit 24.

The model selection unit 22 has the same function as the model selection unit 12. Further, the model library 23 has the same function as the model library 13, and the model calculation unit 24 has the same function as the model calculation unit 14.

Similar to the model selection unit 12, the model selection unit 22 stores the abstraction degree setting and the correspondence relationship information indicating the correspondence relationship with the model set. The model set of correspondence information in the second embodiment includes a driver model, a machine model, and a controller model.

Similar to the model selection unit 12, the model selection unit 22 selects a model set corresponding to the abstraction degree setting from the models stored in the model library 23 based on the abstraction degree setting given from the outside. That is, the model selection unit 22 selects a model suitable for the purpose of the simulation from the model library 23 based on the abstraction degree setting. The model set selected by the model selection unit 22 is a combination of a driver model, a machine model, and a controller model.

The abstraction degree setting is input to the model selection unit 22 by the user of the simulation device 21. In the abstraction degree setting of the second embodiment, one of the above two types of options is specified.

Similar to the model library 13, the model library 23 stores a plurality of driver models having different abstractions and a plurality of machine models having different abstractions. Further, the model library 23 stores a plurality of controller models having different abstractions. Also in the second embodiment, the machine models M1 and M2 are software having a common interface.

In the second embodiment, the case where the controller models stored in the model library 23 are the controller models C1 and C2 will be described. The controller models C1 and C2 are models that simulate the operation of the controller 10A included in the mechanical system 20A. The controller model C1 is a model that simulates the operation of the controller 10A with higher accuracy than the controller model C2. Further, the controller model C2 is a model that simulates the operation of the controller 10A more simply than the controller model C1.

The controller models C1 and C2 are models of the controller 10A included in the mechanical system 20A. Examples of the controller 10A are a programmable logic controller, an industrial personal computer, a servo system controller and the like, as described above. The controller 10A executes a sequence program that turns on the positioning start signal of the head 6 after receiving the positioning operation completion signal of the stage 5.

The controller model C1 accurately simulates the processing corresponding to the sequence program executed by the controller 10A. The controller model C1 executes a sequence program waiting for the positioning operation completion signal of the stage 5 at a constant processing cycle, and when the positioning operation completion signal of the stage 5 is obtained, turns on the positioning start signal of the head 6 in the next processing. Simulate the process of generating a position command.

The controller model C2 simplifies and simulates the process executed by the controller 10A. The controller model C2 simulates only the process of generating the position command of each motor.

The model calculation unit 24 reads the driver model, the machine model, and the controller model corresponding to the model set notified from the model selection unit 22 from the model library 23. The model calculation unit 24 performs a simulation calculation of the motor driver, the machine 1, and the controller 10A by using the read driver model, the machine model, and the controller model.

The model calculation unit 24 calculates the model of the mechanical system 20A including the model set selected from the model library 23. For example, assume that the model of the mechanical system 20A includes a motor model. In this case, the model calculation unit 24 includes a driver model and a motor model corresponding to the X-axis 1a, a driver model and a motor model corresponding to the Y-axis 1b, a driver model and a motor model corresponding to the Z-axis 1c, and a mechanical model. And the controller model are included in the model of the mechanical system 20A.

Since the simulation processing procedure by the simulation device 21 is the same as the simulation processing procedure by the simulation device 11 described with reference to FIG. 3, the description thereof will be omitted.

Next, the detailed operation of the simulation device 21 according to the second embodiment will be described. First, as an abstraction level setting, consider a case where the user selects a simulation for adjusting the parameters of the motor driver of the stage 5.

In this case, the model selection unit 22 includes, for example, three driver models D1 corresponding to the X-axis 1a and Y-axis 1b, a driver model D2 corresponding to the Z-axis 1c, and a machine model M1 based on the abstraction degree setting. Select. Further, the model selection unit 22 selects the controller model C2 based on, for example, the abstraction degree setting. The operation of the simulation device 21 at this time is the same as the operation when the simulation device 11 of the first embodiment selects a simulation for the purpose of adjusting the parameters of the motor driver of the stage 5 as the abstraction degree setting. be.

Next, as an abstraction level setting, consider a case where the user selects a simulation for the purpose of confirming the operation of the sequence program that determines the operation order of the stage 5 and the head 6.

In this case, the model selection unit 22 sets the three driver models D2 corresponding to, for example, the X-axis 1a, the Y-axis 1b, and the Z-axis 1c, the mechanical model M2, and the controller model C1 based on the abstraction degree setting. select.

FIG. 5 is a diagram for explaining a sequence program used in a mechanical system in which the simulation apparatus according to the second embodiment simulates an operation. FIG. 5 shows a procedure of processing executed by the sequence program.

Each process of the blocks BL10, BL20, and BL30 is implemented in the sequence program 25 shown in FIG. In the sequence program 25, it is assumed that the processing of one block is calculated for each calculation cycle of the controller 10A.

When the controller 10A starts executing the sequence program 25, the positioning of the stage 5 is started (block BL10). In the sequence program 25, it is determined whether or not the positioning of the stage 5 is completed (block BL20).

If the positioning of the stage 5 is not completed (blocks BL20, No), the controller 10A does not proceed to the next process and waits until the positioning of the stage 5 is completed. That is, the block BL20 has a specification that does not transition to the processing of the next block until the positioning completion bit of the stage 5 is turned ON.

When the positioning of the stage 5 is completed (block BL20, Yes), the controller 10A starts positioning the head 6 (block BL30).

The controller model C1 simulates the processing of the sequence program 25 shown in FIG. Specifically, when the controller model C1 simulates the processing of the block BL10, that is, when the positioning of the stage 5 is started, a position command is given to the driver models of the X-axis 1a and the Y-axis 1b that drive the stage 5. Be done.

After that, the controller model C1 transitions to the processing of the block BL20. The driver model of the X-axis 1a and the Y-axis 1b is the driver model D2, and the mechanical model of the X-axis 1a and the Y-axis 1b is the mechanical model M2. Further, since the machine models of the X-axis 1a and the Y-axis 1b are the machine model M2, the controller model C1 does not simulate the precise vibration characteristics of the machine 1. In the sequence program 25, the positioning of the stage 5 is not completed until all the position commands to the stage 5 are issued. Therefore, in one calculation cycle, it is determined that the positioning of the stage 5 is completed in the block BL20. Not given.

When the positioning of the stage 5 is completed, the controller model C1 transitions to the processing of the block BL30 in the next calculation cycle of the controller 10A. As a result, the positioning of the head 6 is started.

In this way, the controller model C1 can accurately confirm the operation of the sequence program 25. On the other hand, since the controller model C1 simplifies the operations of the motor driver, the motor, and the machine 1, the calculation cost and the time cost required for the simulation can be kept small.

As described above, in the second embodiment, the user selects one of the purposes of a plurality of types of simulations, for example, two types, and inputs the purpose to the simulation device 21 as an abstraction degree setting. Based on the abstraction degree setting, the model selection unit 22 has an appropriate abstraction degree model such that the required calculation amount is small while having the accuracy required for the simulation from the models stored in the model library 23. Is selected.

For example, when a simulation aimed at adjusting the parameters of the motor driver of the stage 5 is selected, the model selection unit 22 performs a precise driver model D1 and a mechanical model corresponding to the X-axis 1a and the Y-axis 1b that drive the stage 5. Select M1. Since the simulation accuracy of the sequence program 25 does not affect the adjustment of the motor driver, the model selection unit 22 selects the controller model C2 having a low simulation accuracy as the controller model for simulating the sequence program 25.

In general, industrial controllers are often designed to perform high-speed processing in pursuit of real-time performance, and the calculation cycle is short, and enormous calculation costs are required to simulate these operations by simulation. .. Since the simulation device 21 of the second embodiment lowers the simulation accuracy of the controller model for the parameter adjustment that does not require the sequence program 25 to be simulated with high accuracy, the calculation cost of the simulation can be reduced. .. On the other hand, since the simulation device 21 uses a highly accurate model for the motor driver and the mechanical model when adjusting the parameters, it is possible to correctly adjust the parameters of the motor driver. That is, the simulation device 21 can execute a simulation using a model having an appropriate degree of abstraction to satisfy the purpose of the simulation while keeping the calculation cost and the time cost small.

Further, when a simulation for the purpose of confirming the operation of the sequence program 25 that determines the operation order of the stage 5 and the head 6 is selected, the model selection unit 22 uses the sequence program 25 as a controller model that simulates the sequence program 25 with high accuracy. The controller model C1 that simulates the processing of is selected. Further, when the operation of the sequence program 25 is confirmed, the model selection unit 22 selects the driver model D2 and the machine model M2 that perform simplified simulation processing.

In checking the operation of the sequence program 25, it is sufficient if it can be confirmed that the calculation of the controller 10A is correct, and detailed simulation of the vibration characteristics of the machine 1 is unnecessary. The simulation device 21 of the second embodiment can confirm the processing by the sequence program 25 while keeping the calculation cost and the time cost small by simplifying the driver model and the machine model.

As described above, according to the second embodiment, the simulation device 21 aims at the simulation based on the abstraction degree setting regardless of which abstraction degree setting is specified by the user, as in the first embodiment. It is possible to select an appropriate level of abstraction model according to the above. Therefore, the simulation device 21 can easily realize a simulation with high calculation efficiency.

Further, since the simulation device 21 extracts and simulates only a part of the blocks of the sequence program 25, the simulation can be simplified and the operation of the mechanical system 20A can be simulated.

Embodiment 3.
Next, the third embodiment will be described with reference to FIGS. 6 and 7. In the first and second embodiments, the case where the machine model is stored in the model libraries 13 and 23 in advance has been described, but in the third embodiment, the simulation device generates the machine model and stores it in the model library.

FIG. 6 is a diagram showing a configuration of a simulation device according to a third embodiment. Of the components of FIG. 6, components that achieve the same functions as the simulation device 11 of the first embodiment shown in FIG. 2 are designated by the same reference numerals, and redundant description will be omitted.

The mechanical system handled as the target of the simulation in the third embodiment is the same mechanical system 20A as in the first embodiment. That is, the simulation device 31 simulates the operation of the mechanical system 20A in the same manner as the simulation device 11. Also in the third embodiment, one of the two simulation purposes is specified by the abstraction degree setting specified by the user.

The simulation device 31 includes a model selection unit 12, a model library 33, a model calculation unit 14, and a model generation unit 36.

Like the model library 13, the model library 33 holds a plurality of driver models having different abstractions and a plurality of machine models having different abstractions. Specifically, the model library 33 holds the driver models D1 and D2 and the mechanical models M1 to M3. As described above, the model library 33 of the third embodiment holds the machine model M3 generated by the model generation unit 36 in addition to the machine models M1 and M2 prepared in advance.

The model generation unit 36 generates a machine model based on the actual machine information input from the outside. The actual machine information, which is the data of the actual machine, is a signal obtained by actually operating the mechanical system 20A. The actual machine information includes, for example, time-series data of the motor positions of the X-axis 1a, Y-axis 1b, and Z-axis 1c, and time-series data of the stage position and the head position.

The mechanical models M1 to M3 are software having a common interface as in the first embodiment. In the first embodiment, the common interface of the mechanical models M1 to M3 has an input interface capable of acquiring the motor position of the X-axis 1a, the motor position of the Y-axis 1b, and the motor position of the Z-axis 1c. It shall be. Further, it is assumed that the common interface of the mechanical models M1 to M3 has an output interface for outputting information on the head position and the stage position.

The model generation unit 36 generates a machine model M3 that outputs time-series data of the stage position and the head position by inputting the time-series data of the motor positions of the X-axis 1a, the Y-axis 1b, and the Z-axis 1c.

The model generation unit 36 uses a machine learning technique such as a neural network or support vector regression, or a system identification technique such as a prediction error method or a subspace method in order to obtain a relationship between an input and an output. Here, an example in which the model generation unit 36 uses a neural network will be described.

The model generation unit 36 uses a neural network for learning using the time-series data of the motor positions of the X-axis 1a, the Y-axis 1b, and the Z-axis 1c and the time-series data of the stage position and the head position as learning data. Do the action. That is, the model generation unit 36 is a machine learning device. The model generation unit 36 has the functions of a state observation unit, a data acquisition unit, and a learning unit. The state observation unit observes the time-series data of the motor positions of the X-axis 1a, Y-axis 1b, and Z-axis 1c as state variables, and the data acquisition unit acquires the time-series data of the stage position and head position as teacher data. do. The learning unit is the time-series data of the motor positions of the X-axis 1a, Y-axis 1b, and Z-axis 1c output from the state observation unit, and the time-series data of the stage position and head position output from the data acquisition unit. The machine model M3 is trained based on the data set created based on the combination.

The model generation unit 36, which is a machine learning device, may be a device separate from the simulation device 31 connected to the simulation device 31 via a network. Further, the model generation unit 36 may exist on the cloud server.

FIG. 7 is a diagram for explaining a neural network used by the simulation apparatus according to the third embodiment. The neural network has an input layer X1, an intermediate layer Y1, and an output layer Z1. M (M is a natural number) of input data (i1, i2, ..., IM) indicating the motor positions of the X-axis 1a, the Y-axis 1b, and the Z-axis 1c are input to the input layer X1. From the output layer Z1 at the right end, N (N is a natural number) output data (o1, o2, ..., ON) indicating the stage position and the head position are output.

The weighting coefficients for each node of the input layer X1 to each node of the intermediate layer Y1 can be set independently, but in FIG. 7, these weighting coefficients are all represented as the same weighting coefficient W1. Similarly, the weighting coefficients from each node of the intermediate layer Y1 to each node of the output layer Z1 are all expressed as the same weighting coefficient W2.

In the neural network, the weight coefficient W1 is multiplied by the output value of each node of the input layer X1, and the linear combination of the results obtained by the multiplication is input to each node of the intermediate layer Y1. Further, in the neural network, the weight coefficient W2 is multiplied by the output value of each node of the intermediate layer Y1, and the linear combination of the results obtained by the multiplication is input to the node of the output layer Z1. At each node of each layer, the output value may be calculated from the input value by a non-linear function such as a sigmoid function. Further, in the input layer X1 and the output layer Z1, the output value may be a linear combination of the input values.

The model generation unit 36 uses the time-series data of the motor positions of the X-axis 1a, the Y-axis 1b, and the Z-axis 1c, and the time-series data of the stage position and the head position, and uses the weight coefficient W1 and the weight coefficient of the neural network. Calculate with W2. The model generation unit 36 can calculate the weighting coefficient W1 and the weighting coefficient W2 of the neural network by using the backpropagation method or the gradient descent method. However, as long as the calculation method can obtain the weight coefficients W1 and W2 of the neural network, the calculation method of the weight coefficients W1 and W2 by the model generation unit 36 is not limited to the above method.

Once the weighting factor of the neural network was determined, the relational expression between the time-series data of the motor positions of the X-axis 1a, Y-axis 1b, and Z-axis 1c and the time-series data of the stage position and the head position was obtained. become. Up to this point, an example of performing learning using a three-layer neural network has been described, but learning using a neural network by the model generation unit 36 is not limited to the above example.

By the operation of the model generation unit 36 described above, a mechanical model M3 is obtained in which the motor positions of the X-axis 1a, Y-axis 1b, and Z-axis 1c are input and the stage position and head position are output.

The simulation accuracy by the neural network can be arbitrarily set by changing the number of layers or the number of nodes of the neural network. That is, the model generation unit 36 can easily generate machine models having various abstractions to be fitted to the actual machine information by using various mathematical formulas as the learning model. This makes it possible to easily generate a plurality of mechanical models having different accuracy for the simulation device 31. The learning by the model generation unit 36 is the first machine learning.

The model generation unit 36 registers the information in which the abstraction degree setting and the machine model M3 are associated with each other in the correspondence information of the model selection unit 12. As a result, the model selection unit 12 can select an appropriate machine model according to the purpose of the simulation based on the abstraction degree setting and the correspondence information. Further, the model generation unit 36 may notify the user that the machine model M3 has been generated and the abstraction degree setting corresponding to the machine model M3. The model generation unit 36 causes the display device to display, for example, the generation of the machine model M3 and the abstraction degree setting corresponding to the machine model M3.

The model selection unit 12 is not limited to the case of generating the machine model M3, but the machine models M1 to M3 such as the machine model M4 and the machine model M5 having different abstractions by changing the number of layers of the neural network or the number of nodes. You may generate a mechanical model different from. These mechanical models M1 to M5 are generated as software having a common interface.

In the third embodiment, the case where the mechanical model M3 generated by the model generation unit 36 outputs the stage position and the head position has been described, but the output of the mechanical model M3 is not limited to the stage position and the head position. The mechanical model M3 may use, for example, the temperature or sound of the mechanical system 20A as the output of the mechanical system 20A. In this way, the model generation unit 36 can set any of the information included in the actual machine information to the input and output of the machine model M3.

Further, the model generation unit 36 is not limited to the actual machine information, and may use, for example, CAD data output by CAD (Computer Aided Design) software and the actual machine information for generating the machine model M3. The CAD data is data having design information of the machine 1. The model generation unit 36 may generate the machine model M3 based on the design information and the actual machine information. As an example of the method of generating the mechanical model M3 based on the design information and the actual machine information, a method of generating the differential equation of the machine model M3 based on the design information and determining the parameters of the differential equation based on the actual machine information is used. be.

As described above, according to the third embodiment, the model generation unit 36 generates the machine model M3 based on the actual machine information. Therefore, even if the user does not prepare the machine model M3 in advance, the model selection unit 12 can generate the machine model M3. You will be able to select the appropriate model from multiple models with different levels of abstraction. Therefore, even in the third embodiment, a simulation with high calculation efficiency can be easily realized.

Embodiment 4.
Next, the fourth embodiment will be described with reference to FIGS. 8 and 9. In the first embodiment, the case where the machine model is stored in the model library 13 in advance has been described, but in the fourth embodiment, the simulation device changes the abstraction degree of the machine model stored in the model library. Generate a mechanical model.

FIG. 8 is a diagram showing a configuration of a simulation device according to a fourth embodiment. Of the components of FIG. 8, components that achieve the same functions as the simulation device 11 of the first embodiment shown in FIG. 2 are designated by the same reference numerals, and redundant description will be omitted.

The mechanical system handled as the target of the simulation in the fourth embodiment is the same mechanical system 20A as the first embodiment. That is, the simulation device 41 simulates the operation of the mechanical system 20A in the same manner as the simulation device 11. Also in the fourth embodiment, one of the two simulation purposes is specified by the abstraction degree setting specified by the user.

The simulation device 41 includes a model selection unit 12, a model library 43, a model calculation unit 14, and a model simplification unit 47.

The model library 43 holds a plurality of driver models having different abstractions and a plurality of mechanical models having different abstractions. In the fourth embodiment, the case where the driver models D1 and D2 are stored in the model library 43 in advance will be described. Further, the model library 43 holds a machine model M1 that simulates the operation of the machine 1 with high accuracy, and a machine model M6 generated by the model simplification unit 47. The mechanical model M1 is the first mechanical model, and the mechanical model M6 is the second mechanical model.

The model simplification unit 47 generates a new model by simplifying the model read from the model library 43. In the fourth embodiment, a case where the model simplification unit 47 reads the machine model M1 and registers the machine model M6 generated by simplifying the machine model M1 in the model library 43 as a new machine model will be described.

The model simplification unit 47 registers the information in which the abstraction degree setting and the machine model M6 are associated with each other in the correspondence information of the model selection unit 12. As a result, the model selection unit 12 can select an appropriate machine model according to the purpose of the simulation based on the abstraction degree setting and the correspondence information. Further, the model simplification unit 47 may notify the user that the machine model M6 has been generated and the abstraction degree setting corresponding to the machine model M6. The model simplification unit 47 displays, for example, the generation of the machine model M6 on the display device and the abstraction degree setting corresponding to the machine model M6.

The mechanical models M1 and M6 are software having a common interface. For example, it is assumed that the common interface of the mechanical models M1 and M6 has an input interface capable of acquiring the motor position of the X-axis 1a, the motor position of the Y-axis 1b, and the motor position of the Z-axis 1c. Further, it is assumed that the common interface of the mechanical models M1 and M6 has an output interface for outputting information on the stage position and the head position.

FIG. 9 is a diagram for explaining a process in which the simulation apparatus according to the fourth embodiment generates a new model by simplifying the model. In the upper part of FIG. 9, the arithmetic circuit 91 for calculating the stage position and the head position among the arithmetic circuits included in the machine model M1 is shown as a block diagram. In the lower part of FIG. 9, the arithmetic circuit 96 for calculating the stage position and the head position among the arithmetic circuits included in the machine model M6 is shown as a block diagram.

The arithmetic circuit 91 of the machine model M1 includes a block 48a that calculates and outputs the stage position, and a block 48b that calculates and outputs the head position. The X-axis motor position, which is the motor position of the X-axis 1a, the Y-axis motor position, which is the motor position of the Y-axis 1b, and the Z-axis motor position, which is the motor position of the Z-axis 1c, are input to the blocks 48a and 48b. ..

The block 48a calculates and outputs the stage position based on the X-axis motor position, the Y-axis motor position, and the Z-axis motor position. Further, the block 48b calculates and outputs the head position based on the X-axis motor position, the Y-axis motor position, and the Z-axis motor position.

In this way, the arithmetic circuit 91 is a circuit in which the X-axis motor position and the Y-axis motor position affect the head position, and the Z-axis motor position, which is the Z-axis 1c motor position, interferes with the stage position. There is. That is, in the mechanical model M1, even if the Z-axis motor position is fixed, when the X-axis motor position or the Y-axis motor position is suddenly changed, the position of the head 6 also undergoes transient vibration. It means that. Further, in the mechanical model M1, even if the X-axis motor position and the Y-axis motor position are fixed, when the Z-axis motor position is suddenly changed, the position of the stage 5 is also transiently vibrated. It means that. As described above, the mechanical model M1 can be said to be a highly accurate model that simulates interference vibration between orthogonal axes.

The model simplification unit 47 generates a mechanical model M6 simplified so as to ignore the influence of the above-mentioned interference vibration. Specifically, the model simplification unit 47 deletes the term including the Z-axis motor position from the calculation formula for calculating the stage position, and includes the X-axis motor position and the Y-axis motor position from the calculation formula for calculating the head position. Delete the term.

The arithmetic circuit 96 of the mechanical model M6 generated in this manner includes a block 48c that calculates and outputs the stage position, and a block 48d that calculates and outputs the head position. The block 48c is a block that calculates the stage position from the X-axis motor position and the Y-axis motor position. The block 48d is a block that calculates the head position only from the Z-axis motor position.

As described above, in the fourth embodiment, the model simplification unit 47 generates a new machine model M6 that executes the simplified simulation from the machine model M1 that executes the high-precision simulation.

In the fourth embodiment, the model simplification unit 47 performs simplification so as to ignore the interference vibration between the axes, but the simplification method is not limited to this method. The model simplification unit 47 may perform simplification so as to ignore the vibration characteristics when calculating the stage position from the X-axis motor position, for example.

Further, the model simplification unit 47 may perform simplification such that the portion simulating heat generation is ignored from the machine model that simulates heat generation in addition to the vibration characteristics of the machine 1. Further, the model simplification unit 47 may simplify the machine model expressed by the finite element method in which the constituent elements of the machine 1 are flexible bodies to a multi-rigid body model which is a connection of rigid bodies.

Further, in the fourth embodiment, the model simplification unit 47 simplifies the mechanical model, but the model simplification unit 47 may simplify the driver model, the motor model, or the controller model. The model simplification unit 47 performs feedback control only for PID control from, for example, a driver model that performs detailed feedback control by adding a notch filter and a low-pass filter to PID (Proportional-Integral-Differential) control. The driver model may be simplified.

As described above, according to the fourth embodiment, the model simplification unit 47 generates a new machine model M6 having a different degree of abstraction. Therefore, even if the user does not prepare the machine model M6 in advance, the model library 43 can be used. , A plurality of machine models M1 and M6 having different degrees of abstraction can be held. That is, even if the degree of abstraction of the machine model registered by the user is only one type, the model simplification unit 47 can generate a plurality of simple machine models from one type of machine model. This eliminates the need for the user to simplify the mechanical model. In addition, the user can select an appropriate model from the machine models M1 and M6. Therefore, even in the fourth embodiment, a simulation with good calculation efficiency can be easily realized.

Embodiment 5.
Next, the fifth embodiment will be described with reference to FIGS. 10 and 11. In the first embodiment, the model selection unit 12 selects the model set using the correspondence information associated with the model set corresponding to the abstraction degree setting, but in the fifth embodiment, the model selection unit is abstract. Select the model set corresponding to the degree setting based on machine learning.

FIG. 10 is a diagram showing a configuration of a simulation device according to a fifth embodiment. Of the components of FIG. 10, components that achieve the same functions as the simulation device 11 of the first embodiment shown in FIG. 2 are designated by the same reference numerals, and redundant description will be omitted.

The mechanical system handled as the target of the simulation in the fifth embodiment is the same mechanical system 20A as in the first embodiment. That is, the simulation device 51 simulates the operation of the mechanical system 20A in the same manner as the simulation device 11. Also in the fifth embodiment, one of the two simulation purposes is specified by the abstraction degree setting specified by the user.

The simulation device 51 includes a model selection unit 52, a model library 13, and a model calculation unit 14.

The model selection unit 52 selects a model from the models stored in the model library 13 based on the abstraction degree setting input from the outside. At this time, the model selection unit 52 sets a total of four models including three driver models corresponding to the X-axis 1a, the Y-axis 1b, and the Z-axis 1c and one machine model from the model library 13 as a model set. select.

The model selection unit 52 learns the input / output relationship in which the abstraction degree setting is input and the four models corresponding to the abstraction degree setting are output. The learning by the model selection unit 52 is the second machine learning. The model selection unit 52 is a machine learning device having the same configuration as the model generation unit 36. That is, the model selection unit 52 has the functions of a state observation unit, a data acquisition unit, and a learning unit. The state observation unit observes the abstraction degree setting as a state variable, and the data acquisition unit acquires the information of the model set associated with each abstraction degree setting as teacher data. The learning unit is associated with the abstraction level setting and the abstraction level setting based on the abstraction level setting output from the state observation unit and the data set created based on the model set output from the data acquisition unit. Learn the relational expression with the appropriate model.

The model selection unit 52, which is a machine learning device, may be a device separate from the simulation device 51, which is connected to the simulation device 51 via a network. Further, the model selection unit 52 may exist on the cloud server.

The model selection unit 52 uses, for example, a neural network, support vector regression, or the like in order to obtain the relationship between the input and the output. Here, an example in which the model selection unit 52 uses a neural network will be described.

The model selection unit 52 performs machine learning using a neural network using a plurality of abstraction degree settings and information of a model set associated with each of these abstraction degree settings as learning data. It is assumed that the purpose of the simulation is set in the abstraction degree setting of the fifth embodiment. Further, in the fifth embodiment, in order to facilitate machine learning, it is assumed that the information indicating the purpose of the simulation is information that can be divided into several elements.

For example, the information indicating the purpose of the simulation can be divided into the following three elements. The first element, which is an element of the first element, is an axis whose operation is confirmed in the simulation. Examples of the first element are "X-axis 1a", "Y-axis 1b", "Z-axis 1c", or "all axes".

The second element, which is the second element, is an operation to be confirmed in the simulation. An example of the second element is "operation range confirmation", "power consumption confirmation", or "control parameter adjustment".

The third element, which is the third element, is the accuracy required for the simulation. An example of the third element is "high precision" or "low precision".

FIG. 11 is a diagram for explaining a neural network used by the simulation apparatus according to the fifth embodiment. The neural network has an input layer X2, an intermediate layer Y2, and an output layer Z2. Three input data (e1, e2, e3) of the first element, the second element, and the third element are input to the input layer X2. From the rightmost output layer Z2, a total of four output data (m1, m2, m3, m4) consisting of three driver models corresponding to the X-axis 1a, Y-axis 1b, and Z-axis 1c and one mechanical model. ) Is output.

The weighting coefficients for each node of the input layer X2 to each node of the intermediate layer Y2 can be set independently, but in FIG. 11, these weighting coefficients are all represented as the same weighting coefficient W3. Similarly, the weighting coefficients from each node of the intermediate layer Y2 to each node of the output layer Z2 are all expressed as the same weighting coefficient W4.

In the neural network, the weight coefficient W3 is multiplied by the output value of each node of the input layer X2, and the linear combination of the results obtained by the multiplication is input to each node of the intermediate layer Y2. Further, in the neural network, the weight coefficient W4 is multiplied by the output value of each node of the intermediate layer Y2, and the linear combination of the results obtained by the multiplication is input to the node of the output layer Z2. At each node of each layer, the output value may be calculated from the input value by a non-linear function such as a sigmoid function. Further, in the input layer X2 and the output layer Z2, the output value may be a linear combination of the input values.

The model selection unit 52 calculates the weight coefficients W3 and W4 of the neural network using the plurality of abstraction degree settings and the information of the model set associated with each of these abstraction degree settings as learning data. The model selection unit 52 can calculate the weighting coefficients W3 and W4 of the neural network by using the backpropagation method or the gradient descent method. However, as long as the calculation method can obtain the weight coefficients W3 and W4 of the neural network, the calculation method of the weight coefficients W3 and W4 by the model selection unit 52 is not limited to the above method.

Once the weighting factor of the neural network is determined, the relational expression between the abstraction degree setting and the appropriate model associated with the abstraction degree setting is obtained. Up to this point, an example of performing learning using a three-layer neural network has been described, but learning using a neural network by the model selection unit 52 is not limited to the above example.

By the operation of the model selection unit 52 described above, a neural network that outputs four models with the abstraction degree setting as an input can be obtained.

As described above, according to the fifth embodiment, the model selection unit 52 machine-learns an appropriate model set corresponding to the abstraction degree setting, so that an appropriate model set corresponding to the abstraction degree setting newly input by the user can be obtained. It can be selected from the model library 13. Therefore, even in the fifth embodiment, a simulation with high calculation efficiency can be easily realized.

Embodiment 6.
Next, the sixth embodiment will be described with reference to FIGS. 12 and 13. In the first embodiment, the simulation device 11 that simulates the operation of the mechanical system 20A that drives the stage 5 and the head 6 has been described. However, in the sixth embodiment, the simulation device that simulates the operation of the mechanical system that is different from the mechanical system 20A has been described. Will be described. In the sixth embodiment, a simulation device that simulates the operation of a mechanical system that conveys a sheet material by roll-to-roll will be described.

FIG. 12 is a diagram showing a configuration of a mechanical system whose operation is simulated by the simulation apparatus according to the sixth embodiment. The mechanical system 20B includes a machine 2 driven by a motor and a controller 10B for controlling the machine 2. Further, the mechanical system 20B includes a stamp 7 for printing on the sheet material Ws.

In the following description, the two directional axes in the plane parallel to the printing surface of the sheet material Ws by the stamp 7 and orthogonal to each other are defined as the X direction and the Y direction. Further, the direction orthogonal to the X direction and the Y direction is defined as the Z direction.

The machine 2 has a three-axis motor including a winding shaft 2a extending in the X direction, a winding shaft 2b extending in the X direction, and a Z shaft 2c extending in the Z direction.

The machine 2 controls the transfer of the sheet material Ws by the unwinding shaft 2a and the take-up shaft 2b. In the machine 2, the sheet material Ws is conveyed while repeating the feeding operation and the pause of the sheet material Ws, and when the sheet material Ws is paused, the stamp 7 attached to the Z-axis 2c by the operation of the Z-axis 2c is the sheet material. Print processing is performed on Ws.

A motor driver (not shown) is mounted on the mechanical system 20B. In the mechanical system 20B, the controller 10B controls the motor driver, and the motor driver drives the motor of each axis. Examples of the controller 10B are a programmable logic controller, an industrial personal computer, a servo system controller and the like.

The sheet material Ws is wound around the unwinding shaft 2a and the take-up shaft 2b, and moves in the Y direction by rotating the unwinding shaft 2a and the take-up shaft 2b. The stamp 7 moves in the Z direction by the operation of the Z axis 2c, and prints on the sheet material Ws that moves in the Y direction.

The controller 10B controls the position of the sheet material Ws in the Y direction by controlling the operation of the unwinding shaft 2a and the operation of the take-up shaft 2b. Further, the controller 10B controls the position of the stamp 7 attached to the Z-axis 2c in the Z direction by controlling the operation of the Z-axis 2c.

FIG. 13 is a diagram showing a configuration of a simulation device according to a sixth embodiment. Of the components of FIG. 13, components that achieve the same functions as the simulation device 11 of the first embodiment shown in FIG. 2 are designated by the same reference numerals, and redundant description will be omitted.

The simulation device 71 executes the same simulation as the simulation device 11, but the type of mechanical system to be simulated differs between the simulation device 71 and the simulation device 11.

The simulation device 71 simulates the operation of the mechanical system 20B. Also in the sixth embodiment, one of the two simulation purposes is specified by the abstraction degree setting specified by the user.

The simulation device 71 includes a model selection unit 72, a model library 73, and a model calculation unit 74. Comparing the simulation device 71 and the simulation device 11, the simulation device 11 simulates the operation of the mechanical system 20A, whereas the simulation device 71 simulates the operation of the mechanical system 20B.

The abstraction level setting received by the model selection unit 72 of the simulation device 71 is information indicating the abstraction level of the simulation for the mechanical system 20B. The model selection unit 72 selects a model corresponding to the abstraction degree setting in the same manner as the model selection unit 12. The model selection unit 72 notifies the model calculation unit 74 of the selected model set. Further, the model library 73 stores a plurality of driver models having different abstractions and a plurality of machine models having different abstractions, similarly to the model library 13.

In the sixth embodiment, the case where the driver models stored in the model library 73 are the driver models D11 and D12 will be described. Further, a case where the machine models stored in the model library 73 are two machine models M11 and M12 will be described.

The driver models D11 and D12 are models that simulate the operation of the motor driver, similarly to the driver models D1 and D2. The driver model D11 is a model that simulates the operation of the motor driver with higher accuracy than the driver model D12. The driver model D11 simulates the functions of, for example, a feedforward controller, a position feedback controller, a speed feedback controller, and a current controller. Further, the driver model D12 is a model that simulates the operation of the motor driver more simply than the driver model D11. The driver model D12 simulates, for example, only the function of the feedforward controller.

The mechanical models M11 and M12 are the same models as the mechanical models M1 and M2. The machine models M11 and M12 simulate the operation of the machine 2. The machine model M11 is a model that simulates the operation of the machine 2 with higher accuracy than the machine model M12. The machine model M11 simulates, for example, the vibration characteristics of the machine 2. Further, the machine model M12 is a model that simulates the operation of the machine 2 more simply than the machine model M11. The machine model M12 simulates, for example, the printing position of the stamp 7 on the sheet material Ws. Details of the driver models D11 and D12 and the mechanical models M11 and M12 will be described later.

In the sixth embodiment, a case where two types of options are prepared as a simulation will be described. In the sixth embodiment, for example, two types of simulation options are prepared: a simulation for confirming the print position and a simulation for adjusting the servo parameters of the unwinding shaft 2a.

The plurality of machine models M11 and M12 held in the model library 73 are software having a common interface like the machine models M1 and M2 of the first embodiment. In the sixth embodiment, an object that summarizes information such as the sheet position, which is the position of the sheet material Ws, the printing position, the external force applied to the motor, and the heat generated by the machine 2 is defined, and this object is output to the common interface of the machine model. May be.

In the sixth embodiment, the common interface of the mechanical models M11 and M12 has an input interface capable of acquiring the motor position of the unwinding shaft 2a, the motor position of the take-up shaft 2b, and the motor position of the Z-axis 2c. It is assumed that

The machine model M11 is, for example, a model that considers the dynamics of the machine 2 and simulates the tension of the sheet material Ws. The machine model M12, for example, changes the sheet position and the printing position only by kinematics based on the motor position without considering the dynamics of the machine 2, and simulates the operating range of the sheet position and the printing position. The machine model M12 may simulate only one of the sheet position and the print position.

The model calculation unit 74 reads the driver model and the machine model corresponding to the model set notified from the model selection unit 72 from the model library 73. The model calculation unit 74 performs a simulation calculation of the motor driver and the machine 2 by using the read driver model and the machine model.

The model calculation unit 74 calculates the model of the mechanical system 20B including the model selected from the model library 73. For example, assume that the model of the mechanical system 20B includes a motor model and a controller model. The motor model is a model that simulates the operation of the motor, and the controller model is a model that simulates the operation of the controller 10B. In this case, the model calculation unit 74 includes a driver model and a motor model corresponding to the unwinding shaft 2a, a driver model and a motor model corresponding to the take-up shaft 2b, and a driver model and a motor model corresponding to the Z-axis 2c. The model of the mechanical system 20B includes a mechanical model and a controller model that gives a position command of the motor to the motor driver. The model calculation unit 74 may acquire the motor model and the controller model from the model library 73, or may acquire the motor model and the controller model from other than the model library 73. The model calculation unit 74 outputs the simulation result to an external device such as a display device. As a result, the display device displays the simulation result.

Next, the detailed operation of the simulation device 71 according to the sixth embodiment will be described. First, consider a case where the user selects a simulation for the purpose of confirming the print position as an abstraction degree setting.

When a simulation for confirming the print position is selected as the correspondence information of the model selection unit 72, the three driver models D12 corresponding to the unwinding shaft 2a, the take-up shaft 2b, and the Z-axis 2c. And the machine model M12 are predetermined to be selected. In this case, when the abstraction degree setting corresponding to the simulation for the purpose of confirming the print position is specified to the user, the model selection unit 72 sets the unwinding axis 2a and the winding based on the abstraction degree setting and the correspondence relation information. The driver model D12 is selected as the driver model corresponding to the take-axis 2b and the Z-axis 2c. Further, the model selection unit 72 selects the machine model M12 as the machine model.

The model calculation unit 74 uses these selected models, that is, the driver model D12 and the machine model M12, to execute a simulation that simulates the transfer of the sheet material Ws and the operation of the stamp 7. At this time, the model calculation unit 74 inputs the position command of the unwinding shaft 2a generated by the controller model to the driver model D12 of the unwinding shaft 2a. In this case, the driver model D12 simulates and controls the motor position of the unwinding shaft 2a only by feedforward control. The model calculation unit 74 inputs the position command generated by the controller model to the driver model D12 for the take-up shaft 2b and the Z-axis 2c as well as the take-up shaft 2a. In this case, the driver model D12 simulates and controls the motor positions of the take-up shaft 2b and the Z-axis 2c only by feedforward control. As a result, the model calculation unit 74 calculates the motor positions of the unwinding shaft 2a, the take-up shaft 2b, and the Z-axis 2c using the driver model D12.

The model calculation unit 74 calculates the machine model M12 by inputting the motor positions of the unwinding shaft 2a, the take-up shaft 2b, and the Z-axis 2c calculated in this way. Here, since the machine model M12 is selected as the machine model, the transfer position of the sheet material Ws and the position of the stamp 7 are calculated from the motor position of each axis only by kinematics. As a result, the model calculation unit 74 can simulate at which position of the sheet material Ws to be conveyed that the stamp 7 is printed. The model calculation unit 74 outputs the transfer position of the sheet material Ws and the position of the stamp 7 as the result of the simulation.

Next, consider a case where the user selects a simulation for adjusting the servo parameters of the unwinding shaft 2a as the abstraction degree setting.

When a simulation for adjusting the servo parameters of the unwinding shaft 2a is selected for the correspondence information of the model selection unit 72, the driver model is used as the driver model corresponding to the unwinding shaft 2a and the take-up shaft 2b. It is assumed that the selection of D11 is predetermined. When a simulation for adjusting the servo parameters of the unwinding shaft 2a is selected for the correspondence information, the driver model D12 is selected as the driver model corresponding to the Z-axis 2c, and the machine is used as the machine model. It is assumed that the model M11 is selected in advance. In this case, when a simulation for adjusting the servo parameters of the unwinding shaft 2a is specified to the user, the model selection unit 72 sets the unwinding shaft 2a and the unwinding shaft 2a based on the abstraction degree setting and the correspondence information. Two driver models D11 are selected as the driver models corresponding to the axis 2b. Further, the model selection unit 72 selects the driver model D12 as the driver model corresponding to the Z-axis 2c. Further, the model selection unit 72 selects the machine model M11 as the machine model.

The model calculation unit 74 executes a simulation that simulates the operation of the head position using these selected models, that is, the driver models D11 and D12 and the mechanical model M11. At this time, the model calculation unit 74 inputs the position command of the unwinding shaft 2a generated by the controller model to the driver model D11 of the unwinding shaft 2a. In this case, the driver model D11 simulates and controls the motor position by the feedforward controller, the position feedback controller, the speed feedback controller, and the current controller. The machine 2 is equipped with, for example, a PI controller as a position feedback controller and a speed feedback controller. In this case, the respective proportional gain and integral gain are control parameters, and the position control performance of the motor of the unwinding shaft 2a is changed by changing these control parameters.

Since the driver model of the take-up shaft 2b is the driver model D11 as in the take-up shaft 2a, the position control performance of the motor of the take-up shaft 2b is changed based on the control parameters. Further, since the motor driver model of the Z-axis 2c is the driver model D12, the motor position is controlled only by feedforward control.

Further, as the mechanical model, the mechanical model M11 that simulates the tension of the sheet material Ws is selected. Therefore, the tension of the sheet material Ws fluctuates depending on the operation of the motor. Therefore, the model calculation unit 74 outputs the tension of the sheet material Ws and the position of the stamp 7 as a simulation result. As a result, the user can confirm the tension fluctuation of the sheet material Ws. Further, the user can adjust the control parameters of the unwinding shaft 2a and the take-up shaft 2b so that the tension fluctuation of the sheet material Ws does not occur while confirming the tension fluctuation of the sheet material Ws which is the simulation result.

As described above, in the sixth embodiment, the user selects one of the purposes of a plurality of types of simulations, for example, two types, and inputs the purpose to the simulation device 71 as an abstraction degree setting. Based on the abstraction degree setting and the correspondence information, the model selection unit 72 has the accuracy required for the simulation from the models stored in the model library 73, and the required calculation amount is smaller than the predetermined amount. A model with an appropriate level of abstraction is selected.

For example, when a simulation for confirming the print position is selected, the model selection unit 72 selects the mechanical model M12 and the driver model D12. The mechanical model M12 calculates the transfer position of the sheet material Ws and the position of the stamp 7 corresponding to the motor position based only on kinematics. Since the mechanical model M12 does not include the influence of dynamics such as an external force on the motor position, there is no need to simulate controlling the motor position by feedback control. Therefore, when selecting the mechanical model M12, the model selection unit 72 selects the simple driver model D12 having only feedforward control.

As a result, the simulation device 71 does not need to calculate a complicated differential equation related to the dynamics of the machine 2 and the feedback control, and the calculation of the model calculation unit 74 becomes very simple. Therefore, the simulation device 71 can confirm the print position at a very small calculation cost as compared with the case of calculating the differential equation when executing the simulation of a precise model. That is, the simulation device 71 can execute a simulation using a model having an appropriate degree of abstraction to satisfy the purpose of the simulation while keeping the calculation cost small.

On the other hand, when a simulation for adjusting the servo parameters of the unwinding shaft 2a is selected, the model selection unit 72 performs a detailed simulation as a driver model corresponding to the unwinding shaft 2a and the take-up shaft 2b that convey the sheet. Select the driver model D11 to execute. Further, the model selection unit 72 selects the machine model M11 that executes a precise simulation as the machine model. The reason for selecting a detailed model for the take-up shaft 2b is that the take-up shaft 2a and the take-up shaft 2b interfere with each other via the sheet material Ws, so that the operation of the take-up shaft 2b is ignored when adjusting the take-up shaft 2a. Because it cannot be done.

The model selection unit 72 selects the driver model D12, which is a simple driver model, for the driver model of the Z-axis 2c, which does not affect the control performance of the motor drivers of the unwinding shaft 2a and the take-up shaft 2b.

Then, the user adjusts the parameters of the unwinding shaft 2a while checking the tension fluctuation of the sheet material Ws output as the simulation result. In this case, since the simulation device 71 uses a precise model for the operation of the unwinding shaft 2a, the take-up shaft 2b, and the sheet material Ws for the simulation, the user can accurately adjust the parameters. .. The elaborate model in this case is a model with a large amount of calculation, but the simulation device 71 simplifies the operation of the Z-axis 2c, which does not need to be considered for the adjustment of the control parameters of the motor driver of the unwinding shaft 2a. Use a simulated model. Therefore, the simulation device 71 can reduce the amount of calculation for simulating the operation of the Z-axis 2c. That is, the simulation device 71 can execute a simulation using a model having an appropriate degree of abstraction to satisfy the purpose of the simulation while keeping the calculation cost small.

In this way, regardless of which abstraction degree setting is specified by the user, the simulation device 71 can select an appropriate abstraction degree model according to the purpose of the simulation based on the abstraction degree setting. , Simulation with good calculation efficiency can be easily realized.

In the sixth embodiment, the case where the position control based on the position command is performed for the motor driver of the unwinding shaft 2a and the take-up shaft 2b has been described, but even if the speed control is performed by the speed command. Alternatively, torque control may be performed by a torque command. Further, the mechanical system 20B may be a system in which the motor driver of the take-up shaft 2b is speed-controlled and the motor driver of the take-up shaft 2a is torque-controlled.

As described above, according to the sixth embodiment, as in the first to fifth embodiments, it is possible to select an appropriate abstraction degree model according to the purpose of the simulation based on the abstraction degree setting, and thus the calculation efficiency. A good simulation can be easily realized.

Here, the hardware configurations of the simulation devices 11,21,31,41,51,71 will be described. Since the simulation devices 11,21,31,41,51,71 have the same hardware configuration, the hardware configuration of the simulation device 11 will be described here.

FIG. 14 is a diagram showing an example of a hardware configuration that realizes the simulation apparatus according to the first embodiment. The simulation device 11 can be realized by the input device 300, the processor 100, the memory 200, and the output device 400. An example of the processor 100 is a CPU (Central Processing Unit, a central processing unit, a processing unit, an arithmetic unit, a microprocessor, a microcomputer, a DSP (Digital Signal Processor)) or a system LSI (Large Scale Integration). Examples of the memory 200 are RAM (Random Access Memory) and ROM (Read Only Memory).

The simulation device 11 is realized by the processor 100 reading and executing a computer-executable simulation program for executing the operation of the simulation device 11 stored in the memory 200. It can be said that the simulation program, which is a program for executing the operation of the simulation device 11, causes the computer to execute the procedure or method of the simulation device 11.

The simulation program executed by the simulation device 11 has a modular configuration including a model selection unit 12 and a model calculation unit 14, which are loaded on the main storage device and generated on the main storage device. NS.

The input device 300 receives the abstraction degree setting and sends it to the processor 100. The memory 200 stores the model library 13 and the like. The driver models D1 and D2 and the mechanical models M1 and M2 are stored in the model library 13 stored in the memory 200. Further, the memory 200 is used as a temporary memory when the processor 100 executes various processes. The output device 400 outputs the simulation result calculated by the model calculation unit 14 to a monitor or the like.

Here, the simulation devices 11,21,31,41,51,71 are realized by one computer, but the simulation devices 11,21,31,41,51,71 are realized by a plurality of computers. May be good.

For example, the simulation device 31 may be realized by two computers, and the program including the model calculation unit 14 and the program including the model generation unit 36 may be executed by different computers.

Further, for example, the simulation device 41 may be realized by two computers, and the program including the model calculation unit 14 and the program including the model simplification unit 47 may be executed by different computers.

Further, for example, the simulation device 51 may be realized by two computers, and the program including the model calculation unit 14 and the program including the model selection unit 52 may be executed by different computers.

The simulation program may be provided as a computer program product in an installable or executable format file stored on a computer-readable storage medium. Further, the simulation program may be provided to the simulation device 11 via a network such as the Internet. It should be noted that some of the functions of the simulation device 11 may be realized by dedicated hardware such as a dedicated circuit, and some may be realized by software or firmware.

The configuration shown in the above embodiments is an example, and can be combined with another known technique, can be combined with each other, and does not deviate from the gist. It is also possible to omit or change a part of the configuration.

1,2 Machine, 1a X-axis, 1b Y-axis, 1c, 2c Z-axis, 2a unwinding shaft, 2b take-up shaft, 5 stages, 6 heads, 7 stamps, 10A, 10B controller, 11,21,31,41 , 51,71 simulation device, 12,22,52,72 model selection unit, 13,23,33,73 model library, 14,24,74 model calculation unit, 20A, 20B mechanical system, 25 sequence program, 36 model generation Unit, 47 model simplification unit, 48a, 48b, 48c, 48d block, 91,96 arithmetic circuit, 100 processor, 200 memory, 300 input device, 400 output device, C1, C2 controller model, D1, D2, D11, D12 Driver model, M1-M6, M11, M12 machine model, Wp work, Ws sheet material.

Claims (9)

A machine model that simulates the operation of a machine that includes a motor and is driven by the motor, a motor driver that controls the motor, and a controller that controls the motor driver. It is a simulation program that simulates using a driver model that simulates the operation of the motor driver.
The machine is based on the abstraction degree setting which is the information in which the abstraction degree of the simulation when simulating the operation of the machine system is set from the model library holding the machine model candidate and the driver model candidate. A model selection step for selecting the machine model used when simulating the operation of the motor driver and the driver model used when simulating the operation of the motor driver.
A model calculation step that simulates the operation of a mechanical system using the machine model and the driver model selected in the model selection step, and
Let the computer run
When at least one of the machine model candidate and the driver model candidate is plural, and there are a plurality of machine model candidates, the machine model candidate simulates one machine with a different degree of abstraction. When a plurality of mechanical models are included and there are a plurality of candidates for the driver model, the candidate for the driver model includes a plurality of driver models that simulate one motor driver with different abstractions. Yes,
A simulation program characterized by this. When there are a plurality of candidates for the machine model, the plurality of machine models having different abstractions are software having a common interface.
The simulation program according to claim 1. Based on the actual machine information which is a signal obtained by actually operating the machine system, a machine model having a degree of abstraction different from that of the plurality of machine models held in the model library is obtained by the first machine learning. Further, the computer is made to perform the model generation step of generating and holding in the model library.
The simulation program according to claim 1 or 2. In the model generation step, a neural network is used in which the time-series data of the position of the motor is input and the time-series data of the position of the part moved by the motor is output, and the degree of abstraction is different from that of the plurality of machine models. Generate different machine models by the first machine learning,
The simulation program according to claim 3, wherein the simulation program is characterized in that. A model simplification that uses a first machine model, which is one of the plurality of machine models, to generate a second machine model, which is a new machine model with a reduced degree of abstraction of the first machine model. Let the computer perform more abstraction steps,
The simulation program according to any one of claims 1 to 4, wherein the simulation program is characterized in that. Further, the computer is made to perform a learning step of learning the machine model and the driver model selected based on the abstraction degree setting by the second machine learning.
In the model selection step, when a new abstraction degree setting is received, the machine model corresponding to the new abstraction degree setting is generated based on the learning result of the second machine learning by inputting the new abstraction degree setting. Select from the machine model candidates and select the driver model corresponding to the new abstraction setting from the driver model candidates.
The simulation program according to any one of claims 1 to 5, wherein the simulation program is characterized in that. A machine model that simulates the operation of a machine that includes a motor and is driven by the motor, a motor driver that controls the motor, and a controller that controls the motor driver. It is a simulation device that simulates using a driver model that simulates the operation of the motor driver.
The machine is based on the abstraction degree setting which is the information in which the abstraction degree of the simulation when simulating the operation of the machine system is set from the model library holding the machine model candidate and the driver model candidate. A model selection unit that selects the machine model used when simulating the operation of the motor driver and the driver model used when simulating the operation of the motor driver.
A model calculation unit that simulates the operation of a mechanical system using the machine model and the driver model selected by the model selection unit.
With
When at least one of the machine model candidate and the driver model candidate is plural, and there are a plurality of machine model candidates, the machine model candidate simulates one machine with a different degree of abstraction. When a plurality of mechanical models are included and there are a plurality of candidates for the driver model, the candidate for the driver model includes a plurality of driver models that simulate one motor driver with different abstractions. Yes,
A simulation device characterized by this. The model library holds a plurality of controller models that simulate the processing of one sequence program executed by the controller with different abstractions.
The model selection unit selects a controller model to be used when simulating the processing of the sequence program from the model library based on the abstraction degree setting.
The model calculation unit simulates the operation of the mechanical system using the controller model selected by the model selection unit.
The simulation apparatus according to claim 7. A machine model that simulates the operation of a machine that includes a motor and is driven by the motor, a motor driver that controls the motor, and a controller that controls the motor driver. It is a simulation method that simulates using a driver model that simulates the operation of the motor driver.
The machine is based on the abstraction degree setting which is the information in which the abstraction degree of the simulation when simulating the operation of the machine system is set from the model library holding the machine model candidate and the driver model candidate. A model selection step for selecting the machine model used when simulating the operation of the motor driver and the driver model used when simulating the operation of the motor driver.
A model calculation step that simulates the operation of a mechanical system using the machine model and the driver model selected in the model selection step, and
Including
When at least one of the machine model candidate and the driver model candidate is plural, and there are a plurality of machine model candidates, the machine model candidate simulates one machine with a different degree of abstraction. When a plurality of mechanical models are included and there are a plurality of candidates for the driver model, the candidate for the driver model includes a plurality of driver models that simulate one motor driver with different abstractions. Yes,
A simulation method characterized by that.

Images