Parameter prediction system, parameter prediction device, and parameter prediction method

Abstract

Reduce adjustment time and lead time. [Solution] The system comprises a motor 11, a command unit 13 that sends command values ​​related to operation to the motor 11, a gain parameter setting unit 12 that has a plurality of gain parameters to set in the motor 11, a mechanism unit 14 that operates in the motor 11 according to the torque output generated based on the command value and the gain parameter, a storage unit 15 that stores the gain parameter, the torque output of the motor 11, the command value and the state quantity of the mechanism unit 14, and an optimization calculation unit 16 that reads the contents stored in the storage unit 15 and performs an optimization calculation. The storage unit 15 stores the change in the gain parameter, torque output, command value and state quantity each time the gain parameter set in the motor 11 by the gain parameter setting unit 12 is changed, and the optimization calculation unit 16 calculates the optimal gain parameter using a statistical method with respect to the contents stored in the storage unit 15.

Description

Technical Field

[0001] The present disclosure relates to a parameter prediction system, a parameter prediction device, and a parameter prediction method for optimally controlling a motor.

Background Art

[0002] In recent years, regarding the control of the operation of motors, adjustment of control parameters has been carried out under normal operation and load operation conditions.

[0003] Patent Document 1 discloses a control parameter automatic adjustment method that identifies a control target model, determines whether the identified control target model is available through simulation, and automatically adjusts the control parameters by either of two automatic adjustment methods with different availability of the control target model.

[0004] Thereby, when the error included in the output of the control target model is large, avoiding automatically adjusting the control parameters using the control target model as it is, and when the control target model is available, the control parameters can be set quickly by automatic adjustment using simulation.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] In order to operate a motor, which is a power source of production equipment and devices, with high responsiveness and stable torque output, adjustment of control (gain) parameters is necessary. However, there are cases where appropriate set values cannot be calculated by the automatic adjustment function, or it may take time when an inexperienced operator performs the adjustment.

[0007] More specifically, while previous automatic adjustment functions considered optimizing filter characteristics using machine learning, they did not consider optimizing gain parameters. This meant that simultaneous optimization of various gains (FF gain and FB gain) was not possible, resulting in a narrow search range for the optimal solution and a time-consuming search process. Therefore, there was a demand to search for more optimal parameters and to automatically adjust the combination of motors and the mechanisms that operate them, which could not be adjusted automatically before, thereby reducing adjustment man-hours and lead times.

[0008] This disclosure provides a parameter prediction system, a parameter prediction device, and a parameter prediction method that can reduce adjustment man-hours and lead times. [Means for solving the problem]

[0009] The parameter prediction system according to this disclosure is a parameter prediction system for optimal control of a motor, comprising: a motor; a command unit that sends command values ​​related to operation to the motor; a gain parameter setting unit that has a plurality of gain parameters to be set for the motor; a mechanism unit that operates in the motor in accordance with the output of torque generated based on the command value and the gain parameters; a storage unit that stores the gain parameters, the output of the torque of the motor, the command value and the state quantities of the mechanism unit; and an optimization calculation unit that performs an optimization calculation, wherein the storage unit stores the changed gain parameters, the output of the torque, the command value and the state quantities in accordance with the change in the gain parameters set for the motor by the gain parameter setting unit, and the optimization calculation unit performs an optimization calculation to calculate the optimal gain parameters using a statistical method on the contents stored in the storage unit. This allows us to calculate the optimal gain parameter that should be set for the motor.

[0010] Furthermore, the parameter prediction device according to this disclosure is a parameter prediction device for optimal control of a motor, comprising: a motor; a command unit that sends command values ​​related to operation to the motor; a gain parameter setting unit that has a plurality of gain parameters to be set for the motor; a mechanism unit that operates in the motor in accordance with the output of torque generated based on the command value and the gain parameters; a storage unit that stores the gain parameters, the output of the torque of the motor, the command value and the state quantities of the mechanism unit; and an optimization calculation unit that performs an optimization calculation, wherein the storage unit stores the changed gain parameters, the output of the torque, the command value and the state quantities in accordance with the change in the gain parameters set in the motor by the gain parameter setting unit, and the optimization calculation unit performs an optimization calculation to calculate the optimal gain parameters using a statistical method on the contents stored in the storage unit. This allows us to calculate the optimal gain parameter that should be set for the motor.

[0011] Furthermore, the parameter prediction method according to this disclosure involves sending a command value to a motor, setting a gain parameter to the motor, operating a mechanism in the motor according to the torque output generated based on the command value and the gain parameter, storing the gain parameter, the torque output of the motor, the command value, and the state quantities of the mechanism in a storage unit in response to a change in the gain parameter set in the motor, and performing an optimization calculation to calculate the optimal gain parameter using a statistical method with respect to the contents stored in the storage unit. This allows us to calculate the optimal gain parameter that should be set for the motor. [Effects of the Invention]

[0012] This disclosure provides a parameter prediction system that can reduce adjustment man-hours and lead times. [Brief explanation of the drawing]

[0013] [Figure 1]This block diagram shows the configuration of the parameter prediction system. [Figure 2] This is an example of a flowchart illustrating the operation of a parameter prediction system. [Figure 3] This figure shows an example of the operation of the gain parameter combination calculation unit. [Figure 4] This diagram shows the operation of the optimization unit in which it calculates the optimal gain parameters. [Modes for carrying out the invention]

[0014] Embodiment 1 The parameter prediction system 1 according to this embodiment will be described below with reference to the drawings. Figure 1 is a block diagram showing the configuration of the parameter prediction system 1. Although the parameter prediction system 1 is described here as a single system, it may also be a single device.

[0015] As shown in Figure 1, the parameter prediction system 1 comprises a motor 11, a gain parameter setting unit 12 for setting arbitrary gain parameters for the motor 11, a command unit 13 for issuing operation commands to the motor 11, a mechanism unit 14 that operates according to the torque generated by the motor 11, a storage unit 15 for storing set values, command values, and operating state quantities, and an optimization calculation unit 16 for calculating the optimal gain parameters from the information stored in the storage unit 15.

[0016] Motor 11 receives commands from the command unit 13 and has predetermined gain parameters set by the gain parameter setting unit 12. Motor 11 outputs a predetermined driving force, i.e., a predetermined torque, according to these commands and gain parameters, to operate the mechanism unit 14. The torque output information of motor 11 is transmitted to and stored in the storage unit 15.

[0017] The gain parameter setting unit 12 has a plurality of gain parameters. As will be described in detail later, when searching for optimized parameters, the motor 11 repeats the operation of generating torque while sequentially applying the plurality of gain parameters possessed by the gain parameter setting unit 12. Note that the gain parameters set by the gain parameter setting unit 12 for the motor 11 are also transmitted to and stored in the storage unit 15.

[0018] The command unit 13 outputs a command for the motor 11 to perform a predetermined operation. For example, the command output by the command unit 13 to the motor 11 can include content for changing the strength of torque at a predetermined timing. Therefore, hereinafter, this command will be described as a predetermined command value. Note that the command value output by the command unit 13 to the motor 11 is also transmitted to and stored in the storage unit 15.

[0019] The mechanism unit 14 is a mechanism that operates in response to the torque from the motor 11. The mechanism unit 14 outputs a state quantity when operating by the torque of the motor 11. Here, the state quantity is an operation in the mechanism unit 14 and can include information on speed and position when operating according to the command value. As a specific example, since the mechanism unit 14 is an actual mechanism, the state quantity output from the mechanism unit 14 corresponds to the content of the command value by the command unit 13, and at the same time, a value with a difference in the rise of the speed when the mechanism unit 14 operates is output. Note that the mechanism unit 14 outputs the information on the state quantity to the motor 11 and also to the storage unit 15.

[0020] The storage unit 15 stores the gain parameters input from the gain parameter setting unit 12 to the motor 11, the torque output of the motor 11 when these gain parameters are input, the command value input from the command unit 13 to the motor 11, and the actual state quantity output from the mechanism unit 14. Specifically, each time the gain parameter set by the gain parameter setting unit 12 for the motor 11 is changed, the storage unit 15 stores the gain parameter corresponding to the change, the torque output, the command value, and the state quantity.

[0021] The optimization calculation unit 16 calculates an optimal gain parameter from the information stored in the storage unit 15. In particular, the optimization calculation unit 16 calculates an optimal gain parameter so that an optimal operation is performed for the combination of the motor 11 and the mechanism unit 14 to be in an adjusted state. When the optimization of the gain parameter is completed, the optimization calculation unit 16 outputs the optimized gain parameter to the motor 11.

[0022] Although not shown in FIG. 1, the parameter prediction system 1 can be configured to include a gain parameter combination calculation unit. The gain parameter combination calculation unit can prepare and set the necessary gain parameters for optimization for the gain parameter setting unit 12 in the required number of combinations.

[0023] Although not shown in FIG. 1, the parameter prediction system 1 can be provided with a control unit that controls the processing of each unit. Hereinafter, the operation of the parameter prediction system will be described as being performed according to the control of the control unit. Here, it is assumed that the control unit determines whether the gain parameter set from the gain parameter setting unit 12 to the motor 11 is a predetermined set value.

[0024] FIG. 2 is an example of a flowchart of the operation of the parameter prediction system 1. FIG. 3 is a diagram related to preprocessing that is executed before step S1 in the flowchart shown in FIG. 2. Here, first, the preprocessing according to FIG. 3 will be described.

[0025] As shown in FIG. 3, the gain parameter combination calculation unit can prepare the gain parameters in the required number of combinations and set them in the gain parameter setting unit 12.

[0026] Here, there are countless combinations of gain parameters, and performing all of them would require an enormous number of simulations and an enormous amount of time, making it practically impossible. Therefore, the gain parameter combination calculation unit uses statistical methods and past adjustment results to minimize the number of simulations required.

[0027] For example, the gain parameter combination calculation unit can utilize statistical methods such as central composite programming. Alternatively, the gain parameter combination calculation unit can use a method based on past adjustment results, determining the minimum and maximum gain values ​​that tend to be stable based on past adjustment results, and then determining the experimental conditions within that range.

[0028] Furthermore, the gain parameter combination calculation unit can reduce the time spent estimating the moment of inertia ratio, which frequently fails to measure during gain adjustment, by setting the optimization range to plus or minus 20% from the calculated value. Note that this optimization range of plus or minus 20% from the calculated value is just one example; any value can be set based on experience.

[0029] In other words, as a concrete example, as shown in Figure 3, the multiple underlying gains (gain a, gain b, and gain c) each have predetermined minimum and maximum values. Here, the gain parameter combination calculation unit can calculate multiple gain parameters (gains 1 to m in Figure 3) to be set in the gain parameter setting unit 12 by combining these gains based on statistical methods and at least one of past adjustment results.

[0030] Next, we will explain each step according to the flowchart shown in Figure 2.

[0031] The control unit controls the motor 11 to set one gain parameter from the gain parameter setting unit 12, and also controls the operation of the mechanism unit 14 (step S1).

[0032] The memory unit 15 records the gain parameters used in step S1 and the state quantities output by the mechanism unit 14 (step S2). Typically, the memory unit 15 also records the torque output value from the motor 11 and the command value output from the command unit 13, in association with these gain parameters and state quantities.

[0033] The control unit determines whether the gain parameter used has reached a predetermined set value (step S3). If the gain parameter used has not reached the set value (No in step S3), the process returns to step S1 and sets another gain parameter.

[0034] As a specific example, the gain parameter setting unit 12 arranges the gain parameter values ​​themselves in ascending or descending order, etc. The control unit changes the gain parameters set to the motor 11 from the gain parameter setting unit 12 according to this order, and repeats the operations of steps S1 and S2 until the gain parameter values ​​reach predetermined set values. As a result, the storage unit 15 stores information such as command values ​​and state variables along with the changes in the gain parameters.

[0035] On the other hand, if the gain parameter used reaches a preset value (Yes in step S3), the process proceeds to step S4.

[0036] The optimization calculation unit 16 performs an optimization calculation based on the contents stored in the storage unit 15 (step S4). In the storage unit 15, a combination of multiple pieces of information is recorded, where the torque value, state quantity, and command value are associated with one gain parameter set for the motor 11. The optimization calculation unit 16 can perform an optimization calculation based on this combination of multiple pieces of information.

[0037] Referring to Figure 4, a detailed example of the operation in which the optimization calculation unit 16 calculates the optimal gain parameters in step S4 will be described.

[0038] For example, the optimization calculation unit 16 can use the response phase method, which is a statistical method. Furthermore, the optimization calculation unit 16 can optimize the gains so that the difference between the commanded values ​​and the measured values ​​of state variables such as maximum torque, torque amplitude, and speed becomes a desirable characteristic, based on the operating results under various conditions.

[0039] More specifically, as shown in Figure 4, by operating the mechanism 14 in response to multiple gain parameters (gains 1 to m in Figure 4) set in the gain parameter setting unit 12, a state quantity corresponding to each used gain parameter is output as a result of the operation of the mechanism 14. These used gain parameters, state quantities, command values, and torque values ​​are stored in the storage unit 15.

[0040] The optimization calculation unit 16 performs optimization based on these multiple operation results. In this process, statistical methods can be used to derive the optimal gain parameters based on past adjustment results.

[0041] Furthermore, the motor 11 can operate using the optimal gain parameters derived in this way.

[0042] For example, the command value sent from the command unit 13 is a command for the speed of the operation of the mechanism unit 14 over time. The optimization calculation unit 16 can evaluate the difference between this command value and the value of the speed state variable when the mechanism unit 14 operates, and derive the optimal gain parameter.

[0043] Here, the optimization calculation unit 16 can use a waveform correlation function as a function to quantitatively evaluate the difference between the command value and the state variable. More specifically, the waveform correlation function is one of the indicators used to evaluate whether two time-series waveforms are similar, that is, whether they are correlated, and it can evaluate not only the delay between the command speed and the measured speed, but also the maximum speed value at the same time.

[0044] In this way, the optimization unit can reduce the number of target variables for optimization by utilizing the waveform correlation function, thereby reducing the load on the optimization process.

[0045] As a result, the parameter prediction system 1 can search for more appropriate control parameters than conventional automatic adjustments, and can also automatically adjust the gain parameters set for the motor so that the combination of the motor and the mechanism that operates using the torque from the motor is in an optimal state, which was not possible to adjust automatically before.

[0046] In particular, the optimization calculation unit 16 can perform calculations to calculate the optimal gain parameters using statistical methods on the contents stored in the memory unit 15. As a result, the parameter prediction system 1 can calculate the gain parameters more appropriately and with greater robustness compared to methods used to date.

[0047] Therefore, it is possible to provide a parameter prediction system that can reduce adjustment man-hours and lead times.

[0048] It should be noted that the present invention is not limited to the embodiments described above, and can be modified as appropriate without departing from the spirit of the invention. In other words, the above description has been omitted and simplified as appropriate for the sake of clarity, and those skilled in the art can easily change, add, and modify each element of the embodiments within the scope of the present invention. [Explanation of Symbols]

[0049] 1. Parameter prediction system 11 Motor 12 Gain parameter setting section 13 Command Department 14 Mechanism 15 Storage section 16 Optimization Calculation Unit

Claims

1. A parameter prediction system for optimal motor control, Motor and, The motor is equipped with a command unit that sends command values ​​related to its operation, A gain parameter setting unit having multiple gain parameters for setting the motor, The motor includes a mechanism that operates in accordance with the output of torque generated based on the command value and the gain parameter, A storage unit that stores the gain parameter, the torque output of the motor, the command value, and the state quantity of the mechanism, It comprises an optimization calculation unit that performs optimization calculations, The aforementioned storage unit is In response to a change in the gain parameter set for the motor by the gain parameter setting unit, the unit stores the changed gain parameter, the torque output, the command value, and the state quantity. The optimization calculation unit, An optimization operation is performed on the contents stored in the memory unit to calculate the optimal gain parameters using statistical methods. Parameter prediction system.

2. The aforementioned gain parameter setting unit is further provided with a gain parameter combination calculation unit that sets multiple gain parameters, The aforementioned gain parameter combination calculation unit is: Multiple gain parameters are calculated by combining pre-defined gains based on at least one of a statistical method or past adjustment results, and these parameters are set in the gain parameter setting unit. The parameter prediction system according to claim 1.

3. The command value includes a command for speed per unit of time, The optimization calculation unit, The gain parameters are optimized so that the maximum torque, torque amplitude, and speed exhibit desired characteristics between the command value and the state variable. A parameter prediction system according to claim 1 or claim 2.

4. A parameter prediction device for optimal motor control, Motor and, The motor is equipped with a command unit that sends command values ​​related to its operation, A gain parameter setting unit having multiple gain parameters for setting the motor, The motor includes a mechanism that operates in accordance with the output of torque generated based on the command value and the gain parameter, A storage unit that stores the gain parameter, the torque output of the motor, the command value, and the state quantity of the mechanism, It comprises an optimization calculation unit that performs optimization calculations, The aforementioned storage unit is In response to a change in the gain parameter set for the motor by the gain parameter setting unit, the unit stores the changed gain parameter, the torque output, the command value, and the state quantity. The optimization calculation unit, An optimization operation is performed on the contents stored in the memory unit to calculate the optimal gain parameters using statistical methods. Parameter prediction device.

5. Send command values ​​to the motor, A gain parameter is set for the motor, In the motor, the mechanism operates according to the output of the torque generated based on the command value and the gain parameter. In response to a change in the gain parameter set for the motor, the gain parameter, the torque output of the motor, the command value, and the state quantity of the mechanism are stored in the storage unit. An optimization operation is performed on the contents stored in the memory unit to calculate the optimal gain parameters using statistical methods. Parameter prediction method.

Images