Parameter prediction simulator device

Abstract

To calculate an optimal control parameter on a simulation and apply it to an actual machine.SOLUTION: A parameter prediction simulator device includes a simulation part 3 including a simulation command unit 21, a gain setting unit 22 that sets one gain parameter among a plurality of gain parameters, a motor model unit 23 that outputs a torque value by setting the gain parameter, a mechanism model unit 24 that has a plurality of mechanism models, inputs a torque value to one of the plurality of mechanism models, and outputs a state quantity of a position, a speed, and an acceleration, a storage unit 25 that stores the used gain parameter, the mechanism model, and an output state quantity, and an optimization calculation unit 26 that performs an optimization calculation of the contents stored in the storage unit 25. The mechanism model unit 24 outputs a state quantity according to a combination of the mechanism model and the gain parameter, and the optimization calculation unit 26 calculates the optimum gain parameter using a result of repeatedly performing the calculation of the state quantity and sets the optimum gain parameter in an actual machine unit.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a parameter prediction simulator device for optimal control of a motor. [Background technology]

[0002] In recent years, in controlling the operation of a motor, control parameters have been adjusted under normal operating conditions or under load operating conditions.

[0003] Patent Document 1 discloses a method for adjusting control parameters accurately in a short time by using a device equipped with an actual machine section and a simulation section and comparing the state quantities of the actual machine section and the simulation section. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2004-328829 Summary of the Invention [Problem to be solved by the invention]

[0005] This means that if optimal control parameters can be predicted in advance at the design stage, the lead time for equipment production can be shortened. Predicting these control parameters requires a simulator that can accurately reproduce even the variability of the actual equipment. In other words, this will not work unless there is a mechanical model that accurately reproduces the behavior of the actual equipment, and it is difficult to create a model that matches perfectly.

[0006] The present disclosure provides a parameter prediction simulator device that calculates optimal control parameters in a simulation, taking into account differences between an actual machine and a simulation, and applies the calculated control parameters to the actual machine. [Means for solving the problem]

[0007] A parameter prediction simulator device according to the present disclosure is a parameter prediction simulator device that predicts and sets optimal control parameters used in a motor for a real machine unit having an actual machine command unit, a motor that drives based on commands from the actual machine command unit, and a mechanism unit that is connected to the motor and operates according to the driving force of the motor, and includes a simulation command unit that outputs commands, a gain setting unit that sets one of a plurality of gain parameters, a motor model unit in which the gain parameter is set from the gain setting unit and outputs a torque value according to the gain parameter in accordance with a command from the simulation command unit, a mechanism model unit that has a plurality of mechanism models and inputs a torque value from the motor model unit to one of the plurality of mechanism models and outputs state quantities of position, speed, and acceleration, and a gain setting unit that sets one of a plurality of gain parameters from the gain setting unit. and a simulation unit having parameters, a mechanistic model used by the mechanistic model unit, a memory unit that stores the state quantities, and an optimization calculation unit that reads the contents stored in the memory unit and performs optimization calculation, wherein the mechanistic model unit repeatedly outputs a state quantity corresponding to a combination of the one mechanistic model and the one gain parameter while changing the combination of the one mechanistic model and the one gain parameter, and the memory unit stores the combination of the gain parameter and the mechanistic model and the output state quantity, and after the output of the state quantity according to the combination of the gain parameter and the mechanistic model is repeatedly executed and stored in the memory unit, the optimization calculation unit reads the contents stored in the memory unit and performs optimization calculation to calculate optimal gain parameters, and sets the optimal gain parameters to the motor of the actual machine unit. This allows the optimum gain parameters to be calculated, taking into consideration the fact that the actual device and the model used in the simulator do not match. [Effects of the Invention]

[0008] According to the present disclosure, it is possible to provide a parameter prediction simulator device that can calculate optimal control parameters in a simulation and apply them to an actual machine. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a diagram illustrating a configuration and an operation of a parameter prediction simulator device according to a first embodiment. [Figure 2] FIG. 10 is a diagram illustrating calculation of a plurality of gain parameters used in the simulation according to the first embodiment. [Figure 3] FIG. 10 is a diagram illustrating a definition of a range of physical property values ​​according to the first embodiment, a plurality of mechanistic models set using the range of physical property values, and calculation of an optimal gain parameter from an operation simulation result calculated from a combination of these models. DETAILED DESCRIPTION OF THE INVENTION

[0010] Embodiment 1 The parameter prediction simulator device according to this embodiment will be described below with reference to the drawings: Fig. 1(a) shows the configuration of the parameter prediction simulator.

[0011] As shown in FIG. 1( a ), the parameter prediction simulator device 1 includes a real machine unit 2 and a simulation unit 3 .

[0012] The actual machine section 2 includes an actual machine command section 11, a motor 12, and a mechanism section 13.

[0013] The actual machine command unit 11 outputs a command to operate the motor 12 .

[0014] The motor 12 operates in accordance with a command input from the actual machine command unit 11, and outputs a torque value to the mechanism unit 13. The motor 12 operates using a set value input from an optimization calculation unit 26 of the simulation unit 3, which will be described later.

[0015] The mechanism unit 13 is a mechanism that is connected to and operates the motor 12. The mechanism unit 13 outputs a state quantity to the motor 12 in response to the operation of the motor 12.

[0016] The simulation unit 3 includes a simulation command unit 21, a gain setting unit 22 capable of setting a plurality of gain parameters, a motor model unit 23 that receives the gain parameters as input and outputs a torque value, a mechanism model unit 24 that receives the torque value as input and outputs a state, a memory unit 25, and an optimization calculation unit 26.

[0017] The simulation command unit 21 outputs a command to operate the motor model unit 23 .

[0018] The gain setting unit 22 sets one of the plurality of gain parameters for the motor model unit 23. As will be described in detail later, the gain setting unit 22 has an adjustment function of performing a combination calculation for the plurality of set gain values ​​and adjusting the gain parameter to be output to the motor model unit 23.

[0019] Here, adjusting the gain parameters serves the purpose of adjusting equipment parameters to operate the equipment responsively and stably, such as bringing the actual speed, which is delayed due to the influence of inertia and friction compared to the command speed of the equipment, closer to the command value, or preventing torque transmission due to resonance to operate stably.

[0020] The motor model unit 23 uses the gain parameters set by the gain setting unit 22 to output a torque value according to the gain parameters to the mechanism model unit 24.

[0021] The mechanism model unit 24 has multiple mechanism models, and inputs a torque value from the motor model unit 23 to one of the multiple mechanism models, and outputs the state quantities of position, speed, and acceleration to the motor model unit 23 and the memory unit 25.

[0022] As will be described later, the physical property values ​​of the mechanistic models have a predetermined variation. The mechanistic model unit 24 has a function of constructing multiple mechanistic models by randomly selecting the physical property values ​​of the models that are calculated within the predetermined variation.

[0023] The storage unit 25 stores the gain parameters set by the gain setting unit 22 and the state quantities calculated by the mechanism model unit 24 .

[0024] The optimization calculation unit 26 performs optimization by reading a plurality of gain parameters stored in the storage unit 25 and the amount of information calculated corresponding to each of these gain parameters. The optimization calculation unit 26 can set the optimal gain calculated by optimization to the motor 12 of the actual machine unit 2.

[0025] Next, the operation of the parameter prediction simulator device 1 will be described with reference to Fig. 1(b). Fig. 1(b) is a diagram showing a simple operation flow of the parameter prediction simulator device 1.

[0026] Here, it is assumed that the gain setting unit 22 outputs a plurality of gain parameters and a plurality of mechanistic models are constructed in the mechanistic model unit 24. Also, Fig. 2 is a diagram showing a state in which an optimal gain is output by performing optimization from a plurality of operation results obtained by the operation shown in Fig. 1(b) using a plurality of gain parameters and a plurality of mechanistic models.

[0027] As shown in Fig. 1(b), the mechanistic model unit 24 operates the mechanistic model with predetermined gain parameters (step S1). More specifically, one gain parameter (gain parameter x) from among a plurality of gain parameters (gain parameters 1 to m: m is any integer equal to or greater than 2) set in the gain setting unit 22 is set in the motor model unit 23. The motor model unit 23 outputs a torque value based on the set gain parameter to the mechanistic model unit 24. The mechanistic model unit 24 uses the input torque value to calculate a state quantity by assuming that operation is performed using one mechanistic model (mechanistic model y) from among the plurality of mechanistic models (mechanistic models 1 to n: n is any integer equal to or greater than 2).

[0028] The storage unit 25 records the gain parameter x and the mechanical model y used by the mechanical model unit 24 in step S1, and the state quantities calculated using these (step S2). At this time, the mechanical model unit 24 can output the state quantities to the motor model unit 23.

[0029] Here, if the gain parameter x and mechanistic model y used in the calculation have not reached the set values ​​(No in step S3), the processes of steps S1 and S2 are repeated while updating x and y until they reach the set values. On the other hand, if the set x and y have been reached (Yes in step S3), the process proceeds to step S4. In this way, the mechanistic model unit 24 repeatedly calculates the operation results for each combination of the set gain parameters (1 to m) and multiple mechanistic models (1 to n). That is, as shown in FIG. 2, the mechanistic model unit 24 can calculate multiple operation results as the operation simulation results.

[0030] Thereafter, the optimization calculation unit 26 uses a statistical method to optimize a plurality of operation results obtained by combining the gain parameters and the mechanical model stored in the storage unit 25 (step S4). Furthermore, as shown in Fig. 2, the optimization calculation unit 26 can also perform optimization using past adjustment results. As a result, regardless of which mechanical model is used, the optimization calculation unit 26 can determine statistically optimal gain parameters that can operate the mechanism with the desired state quantities (high response, low torque, and low vibration) while taking into account past performance.

[0031] Thereafter, the gain parameters calculated by the optimization calculation unit 26 are set in the motor 12 of the actual unit 2 (step S5).

[0032] From the above, the mechanism model unit 24 outputs state quantities corresponding to combinations of multiple mechanism models and multiple gain parameters set by the gain setting unit 22, the memory unit 25 stores the state quantities output by combinations of gain parameters and mechanism models, and the optimization calculation unit 26 performs optimization and sets the optimized gain parameters to the motor 12 of the actual machine unit 2.

[0033] The parameter prediction simulator device 1 also has a function of calculating a combination of gain parameters in the gain setting unit 22, a function of defining a range of physical property values ​​to be handled in the simulation, and a function of constructing a mechanistic model using this range of physical property values. Details of these three functions will be explained below.

[0034] <Gain parameter combination calculation function> Here, with reference to Fig. 3(a), a description will be given of the combination calculation function in the gain setting unit 22. Fig. 3(a) is a diagram showing a state in which a plurality of gain parameters to be output to the motor model unit 23 are calculated from a plurality of gains registered in the gain setting unit 22.

[0035] 3(a), it is assumed that a plurality of gains, such as a gain, b gain, and c gain, are set in the gain setting unit 22. Here, as an example, it is assumed that the a gain has a minimum value of 1 and a maximum value of 100000, the b gain has a minimum value of 2 and a maximum value of 150000, and the c gain has a minimum value of 10 and a maximum value of 599. It is also assumed that there are other gains besides these.

[0036] The gain setting unit 22 uses a combination calculation function to set multiple gain parameters (1 to m) to be set in the motor model unit 23 using combinations of the set a gain, b gain, c gain, etc., as well as statistical methods and past adjustment results.

[0037] As a result, in the gain setting unit 22, the number of simple combinations of gain a, gain b, gain c, etc. that were set would be extremely large, and running simulations for all combinations would require an enormous number of times and take an enormous amount of time, making it practically impossible. However, the combination calculation function reduces the number of gain parameters to be set in the motor model unit 23, and makes it possible to reduce the number of simulations in the mechanism model unit 24 to the minimum necessary.

[0038] <Function to define the range (variation range) of physical property values ​​to be handled in the simulation> Physical properties have true values. However, to confirm these true values ​​in an actual device, it is necessary to actually operate the device and measure them, but it is difficult to measure the true values ​​of all physical properties.

[0039] As a specific example, physical property values ​​can be identified by measuring the frequency characteristics and torque waveform obtained when two types of commands, a torque command and a speed command, are input to the actual machine, and then allocating the parameters related to this calculation so that the characteristics approach those obtained when the exact same commands are input to the simulation. This makes it possible to reproduce the characteristics of the actual machine, such as resonance, torque oscillation period, and output.

[0040] However, there is variation in the measurement results of actual equipment, making it difficult to uniquely calculate true values ​​in many cases. Furthermore, each physical property has a theoretical value derived from the physical property values ​​listed in the catalog and other general physical property values ​​such as friction, but in reality, theoretical values ​​almost never perfectly match true values. Therefore, even if a simulation is performed using measured physical property values ​​or theoretical values, it is impossible to completely reproduce the behavior of the actual equipment.

[0041] From the above, the simulation unit 3 grasps and uses in advance the range of physical property values ​​that the actual device has, rather than a unique true value or theoretical value.

[0042] 3(b) is a diagram showing a state in which the range of each physical property value is defined by the parameters for which the physical property values ​​have been identified for an axis having a mechanism similar to that of the axis being targeted in the parameter prediction simulator device 1. When physical property value identification is performed for multiple (e.g., four) axes having the same mechanism, the physical property values ​​cannot be uniquely determined to be reproduced in an actual machine, and variations in the physical property values ​​occur. Here, with regard to an axis, typically, one axis is defined by one mechanism consisting of a combination of one motor 12 and a mechanism unit 13.

[0043] 3(b), for each axis there is a theoretical value derived as described above, and a true value that is difficult to identify because it is difficult to measure and variations occur. Therefore, the parameter prediction simulator device 1 obtains the identification results of multiple physical property values ​​for each axis in the actual machine and defines the range of predetermined physical property values, i.e., the range of variation in the physical property values.

[0044] That is, in the mechanistic model unit 24, for the results of physical property values ​​identified on multiple axes, the range between the upper and lower limits of the identified physical property values ​​can be defined as the width of the physical property values ​​on that axis. Note that the theoretical values ​​and true values ​​are included within this width of the physical property values.

[0045] <Function to build a mechanistic model using a range of physical property values> FIG. 3(c) is a diagram showing a state in which the mechanical model unit 24 constructs a plurality of mechanical models using the range of physical property values ​​defined based on the identification results of each axis in the actual machine.

[0046] More specifically, the mechanistic model unit 24 constructs a plurality of mechanistic models (1 to n) using physical property values ​​randomly selected within a defined range of physical property values.

[0047] Note that the physical property values ​​randomly selected within the range of the defined physical property values ​​are typically selected randomly regardless of the identification results of multiple physical property values ​​for each axis in the actual machine described above.

[0048] The optimization calculation unit 26 calculates, as the optimal gain parameters, gain parameters that can be operated stably with high response regardless of the combination of physical property values, i.e., regardless of the mechanistic model having any variation in physical property values. In other words, the optimization calculation unit 26 can calculate, as the optimal gain parameters, gain parameters that reduce the influence of variation in physical property values, regardless of the mechanistic model used.

[0049] Therefore, by using the optimal gain parameters calculated in the optimization calculation unit 26 in this way, it is possible to save the user the trouble of adjusting the gain parameters while operating the actual unit 2. On the other hand, by using the calculated optimal gain parameters, the simulation unit 3 can perform simulation operations with stability and high response, just like the actual unit 2.

[0050] As a result, the parameter prediction simulator device 1 can vary the simulation and calculate an optimal solution, taking into account the fact that the actual machine and the mechanical model used in the simulation do not match. In other words, the parameter prediction simulator device 1 can determine statistically optimal control parameters in the simulation unit 3, taking into account the differences between the actual machine unit 2 and the simulation unit 3.

[0051] As a result, when the parameter prediction simulator device 1 is not used, appropriate gain parameters can be calculated in advance by simulation for the gain parameters being adjusted on the actual machine by using the parameter prediction simulator device 1. Therefore, the lead time required for adjusting the actual machine can be shortened, and the lead time for producing one workpiece on the actual machine can be shortened, thereby improving efficiency.

[0052] The present invention is not limited to the above-described embodiment, and can be appropriately modified without departing from the spirit of the present invention. In other words, the above description has been omitted or simplified as appropriate for the sake of clarity, and a person skilled in the art can easily modify, add, or convert each element of the embodiment within the scope of the present invention.

[0053] Embodiments of the present disclosure may be implemented in hardware or special purpose circuits, software, logic, or any combination thereof. Some aspects may be implemented in hardware, while other aspects may be implemented in firmware or software that may be executed by a controller, microprocessor, or other computing device. [Explanation of symbols]

[0054] 1. Parameter prediction simulator device 2. Actual Machine Section 3 Simulation Section 11. Actual Aircraft Command Center 12 motors 13 Mechanism 21 Simulation Command 22 Gain setting section 23 Motor model section 24 Mechanism Model Section 25 Memory section 26 Optimization calculation unit

Claims

[Claim 1] A parameter prediction simulator device for predicting and setting optimal control parameters used in a motor for a real machine unit having a real machine command unit, a motor that drives based on a command from the real machine command unit, and a mechanism unit that is connected to the motor and operates in response to a driving force of the motor, comprising: a simulation command unit that outputs a command; a gain setting unit that sets one gain parameter among a plurality of gain parameters; a motor model unit to which a gain parameter is set by the gain setting unit and which outputs a torque value according to the gain parameter in response to a command from the simulation command unit; a mechanical model unit having a plurality of mechanical models, the mechanical model unit inputting a torque value from the motor model unit to one of the plurality of mechanical models and outputting state quantities of position, velocity, and acceleration; a storage unit that stores the gain parameters set by the gain setting unit, the mechanistic model used by the mechanistic model unit, and the state quantities; an optimization calculation unit that reads the contents stored in the storage unit and performs optimization calculation; and a simulation unit having the optimization calculation unit; the mechanistic model unit repeatedly outputs a state quantity according to a combination of the one mechanistic model and the one gain parameter while changing the combination of the one mechanistic model and the one gain parameter; the storage unit stores the combination of the gain parameter and the mechanistic model and the output state quantity, the optimization calculation unit repeatedly outputs state quantities by combining the gain parameters and the mechanistic model and stores the results in the storage unit, and then reads the contents stored in the storage unit, performs optimization calculation to calculate optimal gain parameters, and sets the optimal gain parameters in the motor of the actual machine unit. Parameter prediction simulator device.

Images