WO2016042958A1 - Method for identifying plurality of parameters and multiple parameters identification device using method for identifying plurality of parameters - Google Patents

Method for identifying plurality of parameters and multiple parameters identification device using method for identifying plurality of parameters Download PDF

Info

Publication number
WO2016042958A1
WO2016042958A1 PCT/JP2015/072988 JP2015072988W WO2016042958A1 WO 2016042958 A1 WO2016042958 A1 WO 2016042958A1 JP 2015072988 W JP2015072988 W JP 2015072988W WO 2016042958 A1 WO2016042958 A1 WO 2016042958A1
Authority
WO
WIPO (PCT)
Prior art keywords
parameters
parameter
groups
identifying
identification
Prior art date
Application number
PCT/JP2015/072988
Other languages
French (fr)
Japanese (ja)
Inventor
山崎　勝
哲男 梁田
裕理 高野
雄介 上井
Original Assignee
株式会社 日立産機システム
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 株式会社 日立産機システム filed Critical 株式会社 日立産機システム
Publication of WO2016042958A1 publication Critical patent/WO2016042958A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators

Definitions

  • the present invention relates to a model-based development technique for improving the design of a device by grasping the behavior of a target by calculation using a model on a computer, and in particular, using a method for identifying a plurality of parameters and a method for identifying a plurality of parameters.
  • the present invention relates to a plurality of parameter identification devices.
  • test data corresponding to the characteristics of one physical model among N ( ⁇ 2) physical models arranged in parallel is used to determine the physical model and the material parameters of the physical model.
  • a first identification step to identify, a physical model and material parameters identified up to the (n ⁇ 1) th (2 ⁇ n ⁇ N) identification step, and the n ⁇ 1th of the N different physical models Using the test data according to the characteristics of the other physical model that has not been identified until the identification step, the nth physical model and the material parameter of the other physical model are identified. And identifying the material parameters of N different physical models in sequence by performing from the first identification step to the Nth identification step.
  • Patent Document 1 when there is interference with the calculation result by the parameter of the n ⁇ 1th physical model and the parameter of the nth physical model, the interference is caused when the parameter of the nth physical model is changed by identification.
  • the parameter of the (n-1) th physical model that should be originally formed is not corrected, and as a result, the accuracy of parameter identification deteriorates, and the accuracy of the calculation result of the entire calculation model is not sufficient. May end up.
  • an object of the present invention is to provide a method and apparatus capable of realizing high-accuracy parameter identification even when the calculation model includes a number of parameters that require identification and these mutually interfere with the calculation result. There is.
  • a method for identifying a plurality of parameters of a calculation model used for predicting a behavior of a controlled object on a computer the step of grouping a plurality of parameters to be identified into a plurality of groups, and a plurality of parameters for each of a plurality of groups
  • the process of identifying the parameters in the group and the parameter value obtained by identifying the parameters in a group in multiple groups, the calculation of the group that identifies the parameters after the parameters in the group are identified
  • the process of repeating identification of parameters in a plurality of groups for each of a plurality of groups, and the repetition is terminated when a predetermined end condition is satisfied, and the parameter value obtained by the repetition is used as an identification result. And identifying a plurality of parameters.
  • Figure 1 shows an example of a calculation model.
  • the calculation model is composed of two rotary inertias and a shaft connecting them.
  • the 1st rotation inertia 31 is a rotor of an electric motor
  • the 2nd rotation inertia 33 is a flywheel used as load.
  • the shafts to be connected are modeled as torsional characteristics by the rigidity 321 of the shaft and the viscous damping 322 of the shaft.
  • the translational spring and the translational viscous damping symbol are used in the figure for ease of understanding. Actually, the rotational spring and rotational viscous damping are used.
  • Each rotary inertia is provided with a first viscous damping (motor rotor viscous damping) 34 and a second viscous damping (flywheel viscous damping) 35 to account for losses such as friction. These are also viscous damping in the rotational direction.
  • motor rotor inertia J1 flywheel inertia J2
  • shaft rigidity ks shaft viscosity damping coefficient cs
  • motor rotor viscosity damping coefficient c1 motor rotor viscosity damping coefficient c1
  • flywheel viscosity damping coefficient c2 flywheel viscosity damping coefficient c2.
  • the object of the calculation model is a mechanical device driven by an electric motor.
  • FIG. 2 shows an example of a control block diagram including the parameter identification device 1 according to one embodiment of the present invention.
  • a command is supplied from a host controller (not shown), and the command from the host controller is, for example, a target target speed value or target torque value and whether the controller operation mode is set to the speed control mode or torque. This is a combination of control signals for selecting whether to set the control mode.
  • the parameter identification device 1 is upstream of the controller 2, and signals from the motor and machine are input.
  • the output of the parameter identification device 1 is the same signal as the host controller. By switching the switch SW1, it is possible to select which signal the controller 2 receives and operates.
  • the controller 2 receives the input command and signals from the motor 4 and the machine 5 to calculate a target torque and delivers it to the motor amplifier 3.
  • the motor amplifier 3 generates a voltage to be applied to the motor 4 from the target torque and the signal from the motor 4. Although not shown, electric power is supplied from the power source to the motor amplifier 3, and the motor amplifier 3 converts the electric power from the power source and applies it to the motor so that the motor 4 generates a target torque.
  • the motor 4 generates torque from the applied voltage and delivers it to the machine.
  • the machine 5 receives the delivered torque and moves according to the laws of motion including inertial force and frictional force.
  • the signal from the motor 4 to the controller 2 and the motor amplifier 3 is the rotation angle of the motor 4
  • the signal from the machine 5 to the controller 2 is the rotation angle of the flywheel.
  • Each signal is provided with a sensor device that converts a physical quantity into a signal, and is delivered as a sensor output value.
  • FIG. 3 shows an example of processing inside the parameter identification device 1.
  • the measured data storage unit S1 stores, as measured data, signals from the motor 4 and the machine 5 input to the parameter identification device 1 using processing blocks (not shown).
  • a parameter variable of the model is registered and stored in the parameter variable storage unit S2 using a processing block (not shown).
  • the parameter identification unit S4 receives the actual measurement data from the actual measurement data storage unit S1 and the grouped parameters from the parameter variable storage unit S2 through the parameter grouping unit S3.
  • a model execution unit S41 and an identification execution unit S42 are arranged so that actual measurement data and parameter variables input to the parameter identification unit S4 can be input and processed.
  • the parameter identification result calculated by the identification execution unit S42 is output from the parameter identification unit S4 and stored in the parameter value storage unit S5.
  • the model execution unit S41 in the parameter identification unit S4 is preliminarily loaded with a calculation model that can be calculated, and the parameter variable storage unit S2 stores the parameters of the calculation model (J1, J2, ks, cs, c1). , C2) are stored, and the parameter grouping unit S3 is equipped with a distribution process necessary for grouping the parameters.
  • the switch SW1 is connected so that a command from a host controller (not shown) is input to the controller 2, and a command for a target speed value is output from the host controller, for example. Is received by the controller, and the controller 2 generates the target torque so that the motor speed approaches the target speed from the difference between the target speed and the motor speed.
  • the motor amplifier 3 generates a voltage to be applied to the motor 4 from the target torque.
  • the motor 4 of this embodiment is a three-phase synchronous motor, and the motor amplifier 3 generates a three-phase alternating current applied to the motor 4 so that the motor 4 generates a target torque.
  • the motor 4 generates a rotating magnetic field from a three-phase alternating current and generates an electromagnetic torque that rotates the motor rotor.
  • the generated electromagnetic torque rotates the motor rotor, and further rotates the shaft and flywheel that are mechanically connected to the motor rotor.
  • the correction of the target torque by the controller 2 and the correction of the applied voltage and electromagnetic torque are performed every moment, for example, every 1 millisecond of the control cycle, so that the motor rotation is in accordance with the command from the host controller.
  • the operation is controlled.
  • the switch SW1 When performing model parameter identification, the switch SW1 is switched to a connection in which a command from the parameter identification device 1 is input to the controller 2.
  • the model parameter identification device 1 generates a command necessary for performing parameter identification, for example, a rectangular target torque, and supplies it to the controller 2.
  • the controller 2 receives it and generates a target torque by the controller, a voltage to be applied to the motor 4 by the motor amplifier 3 and an electromagnetic torque by the motor 4 in the same manner as in normal operation, and performs a rotational motion.
  • Fig. 4 shows an example of the actual motor rotation speed response at the time of parameter identification.
  • the target torque is a rectangle that gives a constant torque from time 0 seconds to time 1 seconds and becomes 0 from time 1 seconds.
  • the rotation speed response of the motor 4 is a response that rises uniformly from 0 second to 1 second, and falls uniformly after 1 second, and has a shape in which vibration of about 5.5 Hz occurs throughout.
  • the rotational speed response of the motor 4 and the rotational speed response of the flywheel of the machine when performing the rotational motion with the actual machine are read by the parameter identification device 1 and stored in the actual measurement data storage unit S1 as actual measurement data.
  • Fig. 5 shows an example of the flow of parameter identification performed by the parameter identification device.
  • parameters (J1, J2, ks, cs, c1, c2) of the calculation model used for predicting the behavior of the controlled object on the computer are read from the parameter variable recording unit S2, and the parameter group
  • the distribution process of the dividing unit S3 is executed to divide a plurality of parameters into a plurality of groups.
  • the parameters are grouped by physical classification. That is, the group 1 is classified into J1 and J2 which are inertia classes, the group 2 is classified into ks and cs as the class of torsional characteristics in the rotational direction, and the group 3 is classified into three groups c1 and c2 as loss categories.
  • the identification execution unit S42 of the identification parameter identification unit S4 sets the values of J1 and J2 and supplies them to the model execution unit S41, under the same driving conditions as the actual motor rotation speed response and flywheel rotation speed response. Run the model and calculate the model's motor speed response and flywheel speed response.
  • J1 and J2 are corrected so that the rotational speed response of the model and the rotational speed response of the flywheel approach the actual motor rotational speed response and the rotational speed response of the flywheel, and are supplied to the model execution unit S41 again. Calculate the motor speed response of the model and the flywheel speed response.
  • This calculation is repeatedly executed to determine J1 and J2 that are closest to the rotational speed response of the actual machine and the model.
  • initial values registered in advance are set as the values of the parameters.
  • the correction of J1 and J2 is executed so that the evaluation value of the model response is calculated and the evaluation value is minimized.
  • the specific evaluation value was the total sum of squares of the differences between the motor speed response of the actual machine and the model and the rotation speed response of the flywheel.
  • the values of ks and cs are identified by the same calculation as the group 1 parameter identification process.
  • J1 and J2 obtained as a result of the parameter identification process of group 1 immediately before the values of J1 and J2 are used.
  • the parameter identification process F104 of group 3 is executed to determine c1 and c2.
  • a parameter in a plurality of groups is identified for each of a plurality of groups, and a parameter value obtained by identifying a parameter in a certain group in the plurality of groups is changed to a parameter value in a certain group.
  • the parameters in the plurality of groups are identified for each of the plurality of groups while being used for calculation of the group for identifying the parameters after the identification.
  • the end determination unit F105 it is determined whether the parameter value has been sufficiently identified. If it is determined that the predetermined end condition has been achieved, the parameter identification is ended. If the predetermined end condition is not achieved, the process returns to the parameter identification process of group 1 again. In this way, parameter identification for each group is repeated until a predetermined termination condition is achieved, and the finally obtained parameter value is used as the identification result.
  • the predetermined end condition is that the evaluation value of the response of the model is equal to or lower than the predetermined value.
  • the number of repetitions of parameter identification calculation for each group may be a predetermined number or more, or the parameter identification calculation time is a predetermined time or more. But you can.
  • the parameter value obtained as a result of parameter identification is output from the parameter identification unit S4 and stored in the parameter value storage unit S5.
  • the parameter value stored in the parameter storage unit S5 can be taken out by a parameter value output unit (not shown), and can be used in other processes and calculations.
  • the motor rotation speed and the flywheel rotation speed are obtained from the sensor devices attached to the motor rotation speed and the flywheel rotation speed, respectively.
  • the present invention is not limited to this. Good.
  • the driving torque used for identification is a rectangular torque.
  • the present invention is not limited to this, and is not limited to this.
  • a periodic driving torque, a driving torque on a lamp, an impulse driving torque, or an arbitrary driving torque is used.
  • a waveform driving torque may be used.
  • the parameter identification of each group was performed using the same rectangular torque, but the present invention is not limited to this, and the driving torque (operation data) suitable for the identification calculation in the identification of each group.
  • Different drive torques may be used, such as
  • the evaluation value is the sum of squares of the difference between the motor rotation speed response of the actual machine and the model and the rotation speed response of the flywheel, but is not limited to this. It may be the sum of squares of the difference of only the motor speed response, or the Fourier speed of the motor speed response of the actual machine and the model and the speed response of the flywheel, or the rotation speed of one of them, and the resulting peak frequency or The intensity, the average value, the effective value, or the amount of change per unit time may be used.
  • the parameter identification device 1 is arranged upstream of the controller 2, but it may be at the same level as the controller 2 as shown in FIG. 6, or the motor amplifier 3 as shown in FIG.
  • positioned inside may be sufficient.
  • the parameter identification device 1 is arranged off-line, and an operation for acquiring actual machine data necessary for parameter identification is performed using a pre-recorded command.
  • the parameter identification device 1 may be configured to perform parameter identification offline using the recorded data.
  • the group 1 parameter identification process is performed sequentially from the group 1 parameter identification process.
  • the group 1 parameter identification process is executed in parallel with the group 3 parameter identification process.
  • the parameter value gathering process F106 may be a parameter identification flow for integrating the parameter values identified in each group, or as shown in FIG. 10, the parameter identification calculation is performed in parallel with the sequential parameter identification calculation. May be executed in combination, and may be a parameter identification flow for integrating parameter values.
  • Fig. 11 shows an example of the transition of parameter values during parameter identification when the conventional method for identifying all parameters at the same time is used.
  • Fig. 12 shows an example of a comparison of motor rotation speed responses using conventional parameter identification values.
  • FIG. 13 shows an example of a transition diagram of parameter values during parameter identification when parameter identification according to an embodiment of the present invention is used.
  • FIG. 14 shows an example of a comparison diagram of motor rotation speed responses using parameter identification values according to an embodiment of the present invention.
  • parameters are grouped with predetermined characteristics, parameter identification is performed for each group, parameter identification for each group is performed while taking over the result of parameter identification of the immediately preceding group, and each group is identified.
  • the parameters are grouped by physical classification.
  • grouping may be performed using the frequency sensitivity of the parameters.
  • FIG. 15 shows an example of frequency sensitivity analysis of parameters used for this grouping.
  • FIG. 15 is a plot of the result of Fourier transform of the rotational speed of the motor according to the model.
  • the response has a waveform having a peak at the frequency f1.
  • the change of the response is observed by changing each parameter, when the parameters of J1, J2, c1, and c2 are changed, a change in which the peak of the frequency f1 fluctuates is observed.
  • ks and cs are changed, there is no change at the frequency f1, and a change in which a peak appears at the frequency f2 is observed.
  • J1, J2, c1, and c2 are parameters that are sensitive to the frequency f1
  • ks and cs are parameters that are sensitive to the frequency f2. It appears to be.
  • a first group consisting of J1, J2, c1, and c2 and a second group consisting of ks and cs may be grouped. This technique makes it possible to execute parameter grouping at high speed.
  • the parameters are grouped by physical classification.
  • the present invention is not limited to this, and the parameters to be identified may be grouped based on a knowledge database.
  • FIG. 16 is an example of grouping parameters to be identified based on a knowledge database.
  • grouping is performed based on a database that associates responses to be controlled with parameters.
  • the knowledge database S161 As knowledge 1, “the difference in speed response at the time of torque application is due to inertia and friction loss”. According to knowledge 2, “vibration that occurs in the speed waveform during sudden torque changes is due to the torsional characteristics of the shaft” Such knowledge is registered.
  • knowledge database S161 is grouped into first group S162 consisting of J1, J2, c1, and c2 and second group S163 consisting of ks and cs.
  • Fig. 17 shows the relationship between these findings and the speed response.
  • grouping may be performed into a first group consisting of J1, J2, c1, and c2 and a second group consisting of ks and cs. This technique makes it possible to accurately execute parameter grouping based on past knowledge.
  • the parameter identification device 1 is arranged upstream of the controller 2 or at the same level, but it may be at the same level as the motor amplifier 3 as shown in FIG.
  • the parameter identification device of the present invention can take the form of a personal computer connected to the motor amplifier and software operating on the personal computer.
  • a personal computer and a motor amplifier are connected via a communication line, a command is sent from the personal computer to the motor amplifier, and information in the motor amplifier is acquired by the personal computer.
  • Control mode switching, parameter identification processing, data recording, and the like may be performed in cooperation with a personal computer and a motor amplifier.
  • various acquired data and identification results on a personal computer can be taken to another system through another communication line of the personal computer, or the data can be taken to another system by carrying the personal computer itself.
  • the identification calculation of the present invention may be further performed, for example, with higher accuracy to obtain parameters with higher accuracy.
  • the parameter identification device 1 is arranged upstream of the controller 2 or at the same level. However, as shown in FIG. 19, it may be connected to a motor amplifier via a cloud.
  • the parameter identification can be performed by a personal computer or the like installed in the management department. Thereby, a parameter identification device can be installed without being influenced by physical restrictions.
  • parameter identification can be performed on a personal computer in the management department in an integrated manner for a plurality of devices by using the cloud. Further, by monitoring the identification result on a personal computer equipped with a parameter identification device, it is possible to sense a change in the state of the device and to monitor the entire system composed of a plurality of devices.
  • the present invention relates to a model-based development technique for improving the design of a device by grasping the behavior of a target by calculation using a model on a computer, and in particular, with respect to a method and apparatus for identifying a parameter of a model.

Abstract

Provided are a method and a device which are able to achieve high-precision parameter identification even when a calculation model includes a large number of parameters that require identification and these parameters mutually interfere with a calculation result. A method for identifying a plurality of parameters comprises: a step for dividing a plurality of parameters to be identified into a plurality of groups; a step for identifying parameters in the plurality of groups for each of the plurality of groups; a step for repeating the identification of the parameters in the plurality of groups for each of the plurality of groups while carrying over and utilizing a parameter value obtained by the identification of the parameters in a certain group among the plurality of groups for a calculation of a group of which the parameters are identified after the parameters in the certain group were identified; and a step for ending the repetition when a predetermined end condition is achieved and regarding the parameter values obtained by the repetition as identification results.

Description

複数のパラメータを同定する方法および複数のパラメータを同定する方法を用いた複数のパラメータ同定装置A method for identifying a plurality of parameters and a device for identifying a plurality of parameters using a method for identifying a plurality of parameters
　本発明は、計算機上でモデルを用いた計算により対象のふるまいを把握して機器の設計を改善するモデルベース開発技術に関し、特に複数のパラメータを同定する方法および複数のパラメータを同定する方法を用いた複数のパラメータ同定装置に関する。 The present invention relates to a model-based development technique for improving the design of a device by grasping the behavior of a target by calculation using a model on a computer, and in particular, using a method for identifying a plurality of parameters and a method for identifying a plurality of parameters. The present invention relates to a plurality of parameter identification devices.
　本技術分野の背景技術として、特開２０１４－５２３０４号公報（特許文献１）がある。この公報には、並列に配置されたＮ（≧２）個の物理モデルのうちの１つの物理モデルの特徴に応じた試験データを用いて、当該物理モデル、及び、当該物理モデルの材料パラメータを同定する第１の同定ステップと、第ｎ－１（２≦ｎ≦Ｎ）の同定ステップまでに同定された物理モデル及び材料パラメータ、並びに、Ｎ個の異種の物理モデルのうち第ｎ－１の同定ステップまでに同定されていない他の１つの物理モデルの特徴に応じた試験データを用いて、当該他の１つの物理モデル、及び、当該他の１つの物理モデルの材料パラメータを同定する第ｎの同定ステップと、を含み、第１の同定ステップから第Ｎの同定ステップまでを実行することによって、Ｎ個の異種の物理モデルの材料パラメータを順次同定する、と記載されている。 As a background art in this technical field, there is JP 2014-52304 A (Patent Document 1). In this publication, test data corresponding to the characteristics of one physical model among N (≧ 2) physical models arranged in parallel is used to determine the physical model and the material parameters of the physical model. A first identification step to identify, a physical model and material parameters identified up to the (n−1) th (2 ≦ n ≦ N) identification step, and the n−1th of the N different physical models Using the test data according to the characteristics of the other physical model that has not been identified until the identification step, the nth physical model and the material parameter of the other physical model are identified. And identifying the material parameters of N different physical models in sequence by performing from the first identification step to the Nth identification step.
特開２０１４－５２３０４号公報JP 2014-52304 A
　特許文献１では、例えば、第ｎ－１の物理モデルのパラメータと第ｎの物理モデルのパラメータによる計算結果への干渉がある場合、第ｎの物理モデルのパラメータを同定により変更した際に干渉による影響があるので、本来なら成されるべき第ｎ－１の物理モデルのパラメータの修正が成されず、結果としてパラメータの同定の精度が悪化し、計算モデル全体の計算結果の精度が十分ではなくなってしまう場合がある。 In Patent Document 1, for example, when there is interference with the calculation result by the parameter of the n−1th physical model and the parameter of the nth physical model, the interference is caused when the parameter of the nth physical model is changed by identification. As a result, the parameter of the (n-1) th physical model that should be originally formed is not corrected, and as a result, the accuracy of parameter identification deteriorates, and the accuracy of the calculation result of the entire calculation model is not sufficient. May end up.
　そこで、本発明の目的は、計算モデルが多数の同定を必要とするパラメータを含み、これらが相互に計算結果へ干渉する場合であっても高精度なパラメータ同定を実現できる方法及び装置を提供することにある。 Therefore, an object of the present invention is to provide a method and apparatus capable of realizing high-accuracy parameter identification even when the calculation model includes a number of parameters that require identification and these mutually interfere with the calculation result. There is.
　上記課題を解決するために、本発明の特徴は、例えば以下の通りである。 In order to solve the above problems, the features of the present invention are as follows, for example.
　制御対象の挙動を計算機上で予想するために用いる計算モデルの複数のパラメータを同定する方法であって、同定する複数のパラメータを複数のグループにグループ分けする工程と、複数のグループ毎に複数のグループ内のパラメータを同定する工程と、複数のグループ内のあるグループでパラメータが同定されることにより得られたパラメータ値を、あるグループ中のパラメータが同定された以降にパラメータを同定するグループの計算に引き継いで利用しながら、複数のグループ毎に複数のグループ内のパラメータの同定を繰返す工程と、所定の終了条件になったところで繰返しを終了し、繰返しにより得られたパラメータ値を同定結果とする工程と、を有する複数のパラメータを同定する方法。 A method for identifying a plurality of parameters of a calculation model used for predicting a behavior of a controlled object on a computer, the step of grouping a plurality of parameters to be identified into a plurality of groups, and a plurality of parameters for each of a plurality of groups The process of identifying the parameters in the group and the parameter value obtained by identifying the parameters in a group in multiple groups, the calculation of the group that identifies the parameters after the parameters in the group are identified The process of repeating identification of parameters in a plurality of groups for each of a plurality of groups, and the repetition is terminated when a predetermined end condition is satisfied, and the parameter value obtained by the repetition is used as an identification result. And identifying a plurality of parameters.
　本発明によれば、計算モデルが多数の同定を必要とするパラメータを含み、これらが相互に計算結果へ干渉する場合であっても高精度なパラメータ同定を実現できる方法及び装置を提供できる。上記した以外の課題、構成及び効果は、以下の実施形態の説明により明らかにされる。 According to the present invention, it is possible to provide a method and apparatus capable of realizing highly accurate parameter identification even when the calculation model includes a large number of parameters that require identification and these mutually interfere with the calculation result. Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.
計算モデルの例を示す図である。It is a figure which shows the example of a calculation model. 本発明によるパラメータ同定装置を含む制御ブロック図の例である。It is an example of a control block diagram including a parameter identification device according to the present invention. パラメータ同定装置の内部の処理の例を示す図である。It is a figure which shows the example of a process inside a parameter identification device. パラメータ同定時のモータ回転速度応答の例を示す図である。It is a figure which shows the example of the motor rotational speed response at the time of parameter identification. パラメータ同定装置で行われるパラメータ同定のフローの例である。It is an example of the flow of parameter identification performed with a parameter identification apparatus. また別の本発明によるパラメータ同定装置を含む制御ブロック図の例である。It is an example of a control block diagram including another parameter identification device according to the present invention. またさらに別の本発明によるパラメータ同定装置を含む制御ブロック図の例である。It is an example of the control block diagram containing the parameter identification device by another further this invention. パラメータ同定装置をオフラインで実行する場合の制御ブロック図の例である。It is an example of a control block diagram when the parameter identification device is executed offline. また別のパラメータ同定装置で行われるパラメータ同定のフローの例である。It is an example of the flow of parameter identification performed with another parameter identification device. さらにまた別のパラメータ同定装置で行われるパラメータ同定のフローの例である。It is an example of the flow of parameter identification performed with another parameter identification device. 従来のパラメータ同定中のパラメータ値の推移の例である。It is an example of transition of the parameter value during the conventional parameter identification. 従来のパラメータ同定値を用いたモータ回転速度応答の比較の例である。It is an example of the comparison of the motor rotational speed response using the conventional parameter identification value. 本発明によるパラメータ同定中のパラメータ値の推移の例を示す図である。It is a figure which shows the example of transition of the parameter value in the parameter identification by this invention. 本発明によるパラメータ同定値を用いたモータ回転速度応答の比較の例である。It is an example of the comparison of the motor rotational speed response using the parameter identification value by this invention. 同定するパラメータの周波数感度解析の例である。It is an example of the frequency sensitivity analysis of the parameter to identify. 同定するパラメータを知見データベースに基づいてグループ分けする例である。It is an example which groups the parameter to identify based on a knowledge database. 知見と速度応答の関係の例を示す図である。It is a figure which shows the example of the relationship between knowledge and a speed response. モータアンプに接続するパーソナルコンピュータとパーソナルコンピュータ上で動作するソフトウェアの形態の例である。It is an example of the form of the software which operate | moves on the personal computer connected to a motor amplifier and a personal computer. モータアンプとクラウドで接続される形態の例である。It is an example of the form connected with a motor amplifier with a cloud.
　以下、本発明の実施例として、電気モータを用いてフライホイールを回転させる装置のモデルのパラメータ同定に本発明のパラメータ同定装置を用いた例について、図１から図１６を用いて説明する。まず、実施例の構成を説明する。 Hereinafter, as an embodiment of the present invention, an example in which the parameter identification device of the present invention is used for parameter identification of a model of a device that rotates a flywheel using an electric motor will be described with reference to FIGS. First, the configuration of the embodiment will be described.
　図１に計算モデルの例を示す。計算モデルは回転イナーシャ２つとこれらを連結するシャフトから構成されている。ここで、第一の回転イナーシャ３１は電気モータのロータであり、第二の回転イナーシャ３３は負荷となるフライホイールである。連結するシャフトは、シャフトの剛性３２１とシャフトの粘性減衰３２２でねじり特性としてモデル化している。なお、図中では並進運動のばね、並進運動の粘性減衰の記号を用いているが、これは理解をたやすくするためであり、実際は回転方向のばねと回転方向の粘性減衰である。 Figure 1 shows an example of a calculation model. The calculation model is composed of two rotary inertias and a shaft connecting them. Here, the 1st rotation inertia 31 is a rotor of an electric motor, and the 2nd rotation inertia 33 is a flywheel used as load. The shafts to be connected are modeled as torsional characteristics by the rigidity 321 of the shaft and the viscous damping 322 of the shaft. It should be noted that the translational spring and the translational viscous damping symbol are used in the figure for ease of understanding. Actually, the rotational spring and rotational viscous damping are used.
　各回転イナーシャには摩擦などの損失を計上するため、第一の粘性減衰（モータロータの粘性減衰）３４と第二の粘性減衰（フライホイールの粘性減衰）３５を取り付けてある。これらも同じく回転方向の粘性減衰である。ここで、計算モデルのパラメータは、モータロータのイナーシャＪ１、フライホイールのイナーシャＪ２、シャフトの剛性ｋｓ、シャフトの粘性減衰係数ｃｓ、モータロータの粘性減衰係数ｃ１、フライホイールの粘性減衰係数ｃ２の６個となる。本実施例では、計算モデルの対象が電気モータで駆動される機械装置である。 Each rotary inertia is provided with a first viscous damping (motor rotor viscous damping) 34 and a second viscous damping (flywheel viscous damping) 35 to account for losses such as friction. These are also viscous damping in the rotational direction. Here, there are six parameters of the calculation model: motor rotor inertia J1, flywheel inertia J2, shaft rigidity ks, shaft viscosity damping coefficient cs, motor rotor viscosity damping coefficient c1, and flywheel viscosity damping coefficient c2. Become. In this embodiment, the object of the calculation model is a mechanical device driven by an electric motor.
　図２に本発明の一実施形態によるパラメータ同定装置１を含む制御ブロック図の例を示す。図示しない上位の制御器から指令が供給されていて、上位の制御器からの指令は、例えば、目標とする目標速度値や目標トルク値と、制御器の動作モードを速度制御モードとするかトルク制御モードとするかを選択する制御信号を合わせたものである。 FIG. 2 shows an example of a control block diagram including the parameter identification device 1 according to one embodiment of the present invention. A command is supplied from a host controller (not shown), and the command from the host controller is, for example, a target target speed value or target torque value and whether the controller operation mode is set to the speed control mode or torque. This is a combination of control signals for selecting whether to set the control mode.
　パラメータ同定装置１が制御器２の上流にあり、モータ及び機械からの信号が入力されている。パラメータ同定装置１の出力は上位の制御器と同様の信号である。スイッチＳＷ１を切り換えることで、制御器２がどちらの信号を受けて動作するかを選択できるようになっている。 The parameter identification device 1 is upstream of the controller 2, and signals from the motor and machine are input. The output of the parameter identification device 1 is the same signal as the host controller. By switching the switch SW1, it is possible to select which signal the controller 2 receives and operates.
　制御器２は入力された指令とモータ４及び機械５からの信号を受けて目標トルクを算出しモータアンプ３に引き渡す。 The controller 2 receives the input command and signals from the motor 4 and the machine 5 to calculate a target torque and delivers it to the motor amplifier 3.
　モータアンプ３は目標トルクとモータ４からの信号からモータ４に印加する電圧を生成する。図示していないが、電源からモータアンプ３に電力が供給されており、モータアンプ３はモータ４が目標のトルクを発生するように電源からの電力を変換してモータに印加する。 The motor amplifier 3 generates a voltage to be applied to the motor 4 from the target torque and the signal from the motor 4. Although not shown, electric power is supplied from the power source to the motor amplifier 3, and the motor amplifier 3 converts the electric power from the power source and applies it to the motor so that the motor 4 generates a target torque.
　モータ４は印加された電圧から、トルクを生成し機械に引き渡す。 The motor 4 generates torque from the applied voltage and delivers it to the machine.
　機械５は引き渡されたトルクを受けて慣性力や摩擦力などから成る運動の法則に従った運動を行う。ここで、モータ４から制御器２とモータアンプ３への信号は、モータ４の回転角度であり、機械５から制御器２への信号は、フライホイールの回転角度である。それぞれの信号は、それぞれに物理量から信号へ変換するセンサ装置が取り付けられていて、センサ出力値として引き渡される。 The machine 5 receives the delivered torque and moves according to the laws of motion including inertial force and frictional force. Here, the signal from the motor 4 to the controller 2 and the motor amplifier 3 is the rotation angle of the motor 4, and the signal from the machine 5 to the controller 2 is the rotation angle of the flywheel. Each signal is provided with a sensor device that converts a physical quantity into a signal, and is delivered as a sensor output value.
　図３にパラメータ同定装置１の内部の処理の例を示す。実測データ記憶部Ｓ１には図示しない処理ブロックを用いてパラメータ同定装置１に入力されたモータ４及び機械５からの信号などを実測データとして記憶している。パラメータ変数記憶部Ｓ２には図示しない処理ブロックを用いてモデルのパラメータ変数を登録し、これを記憶している。パラメータ同定部Ｓ４には、実測データ記憶部Ｓ１からの実測データ、またパラメータ変数記憶部Ｓ２からパラメータグループ分け部Ｓ３を通してグループ化されたパラメータが入力される。 FIG. 3 shows an example of processing inside the parameter identification device 1. The measured data storage unit S1 stores, as measured data, signals from the motor 4 and the machine 5 input to the parameter identification device 1 using processing blocks (not shown). A parameter variable of the model is registered and stored in the parameter variable storage unit S2 using a processing block (not shown). The parameter identification unit S4 receives the actual measurement data from the actual measurement data storage unit S1 and the grouped parameters from the parameter variable storage unit S2 through the parameter grouping unit S3.
　パラメータ同定部Ｓ４内部には、モデル実行部Ｓ４１と同定実行部Ｓ４２が配置され、パラメータ同定部Ｓ４に入力された実測データやパラメータ変数が入力され処理できるようになっている。同定実行部Ｓ４２で算出されたパラメータ同定結果がパラメータ同定部Ｓ４から出力され、パラメータ値記憶部Ｓ５に格納される構成となっている。 In the parameter identification unit S4, a model execution unit S41 and an identification execution unit S42 are arranged so that actual measurement data and parameter variables input to the parameter identification unit S4 can be input and processed. The parameter identification result calculated by the identification execution unit S42 is output from the parameter identification unit S4 and stored in the parameter value storage unit S5.
　続いて、本実施例の電気モータを用いてフライホイールを回転させる装置の動作を以下に説明する。 Subsequently, the operation of the apparatus for rotating the flywheel using the electric motor of this embodiment will be described below.
　なお、パラメータ同定部Ｓ４内のモデル実行部Ｓ４１には予め計算モデルが計算可能な状態で搭載されていて、パラメータ変数記憶部Ｓ２には、計算モデルのパラメータ（Ｊ１、Ｊ２、ｋｓ、ｃｓ、ｃ１、ｃ２）が格納されており、パラメータグループ分け部Ｓ３では、パラメータをグループ分けするために必要な振り分け処理が搭載されているものとする。 The model execution unit S41 in the parameter identification unit S4 is preliminarily loaded with a calculation model that can be calculated, and the parameter variable storage unit S2 stores the parameters of the calculation model (J1, J2, ks, cs, c1). , C2) are stored, and the parameter grouping unit S3 is equipped with a distribution process necessary for grouping the parameters.
　通常の運転時はスイッチＳＷ１が図示しない上位の制御器からの指令が制御器２に入力されるように接続されており、上位の制御器から、例えば速度目標値の指令が出力され、その指令を制御器が２受け取り、制御器２で目標速度とモータ速度の差からモータ速度が目標速度に近づくように目標トルクを生成する。 During normal operation, the switch SW1 is connected so that a command from a host controller (not shown) is input to the controller 2, and a command for a target speed value is output from the host controller, for example. Is received by the controller, and the controller 2 generates the target torque so that the motor speed approaches the target speed from the difference between the target speed and the motor speed.
　モータアンプ３はその目標トルクからモータ４に印加する電圧を生成する。ここで、本実施例のモータ４は三相同期モータであり、モータアンプ３はモータ４が目標トルクを発生するようにモータ４に印加する三相交流を生成する。 The motor amplifier 3 generates a voltage to be applied to the motor 4 from the target torque. Here, the motor 4 of this embodiment is a three-phase synchronous motor, and the motor amplifier 3 generates a three-phase alternating current applied to the motor 4 so that the motor 4 generates a target torque.
　モータ４は三相交流より回転磁界を発生させ、モータロータを回転させる電磁トルクを発生させる。発生した電磁トルクはモータロータを回転させ、さらにモータロータと機械的に接続されているシャフト、フライホイールを回転させる。そして、制御器２による目標トルクの修正とそれに伴う印加電圧、電磁トルクの修正が時々刻々、例えば制御周期１ミリ秒毎に実施され、上位の制御器からの指令どおりのモータ回転となるように制御される動作となる。 The motor 4 generates a rotating magnetic field from a three-phase alternating current and generates an electromagnetic torque that rotates the motor rotor. The generated electromagnetic torque rotates the motor rotor, and further rotates the shaft and flywheel that are mechanically connected to the motor rotor. Then, the correction of the target torque by the controller 2 and the correction of the applied voltage and electromagnetic torque are performed every moment, for example, every 1 millisecond of the control cycle, so that the motor rotation is in accordance with the command from the host controller. The operation is controlled.
　モデルのパラメータ同定を行う場合は、スイッチＳＷ１をパラメータ同定装置１からの指令が制御器２に入力される接続に切り換える。モデルパラメータ同定装置１はパラメータ同定を行うために必要な指令、例えば、矩形の目標トルクを生成し制御器２に与える。
それを制御器２が受け取り通常の運転時と同様に制御器で目標トルク、モータアンプ３でモータ４に印加する電圧、モータ４で電磁トルクを発生させ、回転運動を行わせる。 When performing model parameter identification, the switch SW1 is switched to a connection in which a command from the parameter identification device 1 is input to the controller 2. The model parameter identification device 1 generates a command necessary for performing parameter identification, for example, a rectangular target torque, and supplies it to the controller 2.
The controller 2 receives it and generates a target torque by the controller, a voltage to be applied to the motor 4 by the motor amplifier 3 and an electromagnetic torque by the motor 4 in the same manner as in normal operation, and performs a rotational motion.
　図４にパラメータ同定時の実機のモータ回転速度応答の例を示す。本例では目標トルクは時刻０秒から時刻１秒の間に一定のトルクを与え、時刻１秒からは０とする矩形とした。モータ４の回転速度応答は、０秒から１秒の間が一様に上昇し、１秒以降が一様に下降する応答となり、全体を通じて約５．５Ｈｚの振動が生じる形状となっている。 Fig. 4 shows an example of the actual motor rotation speed response at the time of parameter identification. In this example, the target torque is a rectangle that gives a constant torque from time 0 seconds to time 1 seconds and becomes 0 from time 1 seconds. The rotation speed response of the motor 4 is a response that rises uniformly from 0 second to 1 second, and falls uniformly after 1 second, and has a shape in which vibration of about 5.5 Hz occurs throughout.
　このようにして実機で回転運動を行った際のモータ４の回転速度応答および機械のフライホイールの回転速度応答をパラメータ同定装置１で読み込み、実測データとして実測データ記憶部Ｓ１に格納する。 Thus, the rotational speed response of the motor 4 and the rotational speed response of the flywheel of the machine when performing the rotational motion with the actual machine are read by the parameter identification device 1 and stored in the actual measurement data storage unit S1 as actual measurement data.
　図５にパラメータ同定装置で行われるパラメータ同定のフローの例を示す。 Fig. 5 shows an example of the flow of parameter identification performed by the parameter identification device.
　まず、パラメータグループ分け処理Ｆ１０１で、パラメータ変数記録部Ｓ２から制御対象の挙動を計算機上で予想するために用いる計算モデルのパラメータ（Ｊ１、Ｊ２、ｋｓ、ｃｓ、ｃ１、ｃ２）を読み出し、パラメータグループ分け部Ｓ３の振り分け処理を実行し、複数のパラメータを複数のグループ分けする。本実施例の処理ではパラメータを物理的な分類でグループ分けした。すなわち、グループ１としてイナーシャの部類であるＪ１、Ｊ２、グループ２として回転方向のねじれ特性の部類であるｋｓ、ｃｓ、グループ３として損失の部類であるｃ１、ｃ２の３グループとした。 First, in the parameter grouping process F101, parameters (J1, J2, ks, cs, c1, c2) of the calculation model used for predicting the behavior of the controlled object on the computer are read from the parameter variable recording unit S2, and the parameter group The distribution process of the dividing unit S3 is executed to divide a plurality of parameters into a plurality of groups. In the processing of this embodiment, the parameters are grouped by physical classification. That is, the group 1 is classified into J1 and J2 which are inertia classes, the group 2 is classified into ks and cs as the class of torsional characteristics in the rotational direction, and the group 3 is classified into three groups c1 and c2 as loss categories.
　次に、グループ１のパラメータ同定処理Ｆ１０２で、Ｊ１、Ｊ２のパラメータ同定を実施する。同定パラメータ同定部Ｓ４の同定実行部Ｓ４２でＪ１とＪ２の値を設定してモデル実行部Ｓ４１に供給し、実機のモータ回転速度応答とフライホイールの回転速度応答を取得した駆動条件と同じ条件でモデルを実行し、モデルのモータ回転速度応答とフライホイールの回転速度応答を計算する。同定実行部Ｓ４２でモデルの回転速度応答とフライホイールの回転速度応答が実機のモータ回転速度応答とフライホイールの回転速度応答に近づくようにＪ１とＪ２を修正して再度モデル実行部Ｓ４１に供給し、モデルのモータ回転速度応答とフライホイールの回転速度応答を計算する。この計算を繰り返し実行し、最も実機とモデルの回転速度応答が近づくＪ１とＪ２を決定する。なお、同定処理の初期では、各パラメータの値は予め登録された初期値が設定されている。また、Ｊ１とＪ２の修正は、モデルの応答の評価値を算出してその評価値が最小となるように実行される。具体的な評価値は実機とモデルのモータ回転速度応答とフライホイールの回転速度応答それぞれの差の二乗和の合計値とした。 Next, in the parameter identification processing F102 of group 1, parameter identification of J1 and J2 is performed. The identification execution unit S42 of the identification parameter identification unit S4 sets the values of J1 and J2 and supplies them to the model execution unit S41, under the same driving conditions as the actual motor rotation speed response and flywheel rotation speed response. Run the model and calculate the model's motor speed response and flywheel speed response. In the identification execution unit S42, J1 and J2 are corrected so that the rotational speed response of the model and the rotational speed response of the flywheel approach the actual motor rotational speed response and the rotational speed response of the flywheel, and are supplied to the model execution unit S41 again. Calculate the motor speed response of the model and the flywheel speed response. This calculation is repeatedly executed to determine J1 and J2 that are closest to the rotational speed response of the actual machine and the model. In the initial stage of the identification process, initial values registered in advance are set as the values of the parameters. The correction of J1 and J2 is executed so that the evaluation value of the model response is calculated and the evaluation value is minimized. The specific evaluation value was the total sum of squares of the differences between the motor speed response of the actual machine and the model and the rotation speed response of the flywheel.
　次に、グループ２パラメータ同定処理Ｆ１０３で、グループ１のパラメータ同定処理と同様の計算でｋｓとｃｓの値を同定する。ここで、グループ２のパラメータ同定処理では、Ｊ１とＪ２の値が直前のグループ１のパラメータ同定処理の結果、得られたＪ１とＪ２を用いるようにする。 Next, in the group 2 parameter identification process F103, the values of ks and cs are identified by the same calculation as the group 1 parameter identification process. Here, in the parameter identification process of group 2, J1 and J2 obtained as a result of the parameter identification process of group 1 immediately before the values of J1 and J2 are used.
　以下同様に、グループ３のパラメータ同定処理Ｆ１０４を実行し、ｃ１とｃ２を決定する。Ｆ１０２、Ｆ１０３、Ｆ１０４では、複数のグループ毎に複数のグループ内のパラメータを同定し、複数のグループ内のあるグループでパラメータが同定されることにより得られたパラメータ値を、あるグループ中のパラメータが同定された以降にパラメータを同定するグループの計算に引き継いで利用しながら、複数のグループ毎に複数のグループ内のパラメータを同定している。 In the same manner, the parameter identification process F104 of group 3 is executed to determine c1 and c2. In F102, F103, and F104, a parameter in a plurality of groups is identified for each of a plurality of groups, and a parameter value obtained by identifying a parameter in a certain group in the plurality of groups is changed to a parameter value in a certain group. The parameters in the plurality of groups are identified for each of the plurality of groups while being used for calculation of the group for identifying the parameters after the identification.
　次に、終了判断部Ｆ１０５で、パラメータ値の同定が十分に実施された状態であるか判断し、所定の終了条件が達成されたと判定された場合はパラメータ同定を終了する。所定の終了条件が達成されていない場合は、再びグループ１のパラメータ同定処理に戻る。このようにして、所定の終了条件が達成されるまで、グループ毎のパラメータ同定を繰り返し、最終的に得られたパラメータ値を同定結果とする。 Next, in the end determination unit F105, it is determined whether the parameter value has been sufficiently identified. If it is determined that the predetermined end condition has been achieved, the parameter identification is ended. If the predetermined end condition is not achieved, the process returns to the parameter identification process of group 1 again. In this way, parameter identification for each group is repeated until a predetermined termination condition is achieved, and the finally obtained parameter value is used as the identification result.
　なお、本実施例では、所定の終了条件はモデルの応答の評価値が所定の値以下となることとしたが、これに限定するものでは無く、所定の終了条件はパラメータ値の変化量が所定の値以下になった場合でもよいし、一巡のグループ毎のパラメータ同定の計算の繰り返し回数が所定の回数以上になった場合でもよいし、パラメータ同定の計算時間が所定の時間以上になった場合でもよい。 In the present embodiment, the predetermined end condition is that the evaluation value of the response of the model is equal to or lower than the predetermined value. The number of repetitions of parameter identification calculation for each group may be a predetermined number or more, or the parameter identification calculation time is a predetermined time or more. But you can.
　次に、パラメータ同定の結果得られたパラメータ値はパラメータ同定部Ｓ４から出力されパラメータ値記憶部Ｓ５に格納される。パラメータ記憶部Ｓ５に格納されたパラメータ値は図示しないパラメータ値出力部によって取り出すことができ、他の処理や計算等で活用することができる。 Next, the parameter value obtained as a result of parameter identification is output from the parameter identification unit S4 and stored in the parameter value storage unit S5. The parameter value stored in the parameter storage unit S5 can be taken out by a parameter value output unit (not shown), and can be used in other processes and calculations.
　なお、本実施例では、モータ回転速度とフライホイールの回転速度をそれぞれに取り付けられているセンサ装置から得られるものとしたが、これに限定するものでは無く、計算で求める推定値であってもよい。 In this embodiment, the motor rotation speed and the flywheel rotation speed are obtained from the sensor devices attached to the motor rotation speed and the flywheel rotation speed, respectively. However, the present invention is not limited to this. Good.
　なお、本実施例では、同定の用いる駆動トルクを矩形トルクとしたが、これに限定するものでは無く、周期的な駆動トルクや、ランプ上の駆動トルク、インパルス的な駆動トルク、また、任意の波形の駆動トルクを用いてもよい。 In the present embodiment, the driving torque used for identification is a rectangular torque. However, the present invention is not limited to this, and is not limited to this. A periodic driving torque, a driving torque on a lamp, an impulse driving torque, or an arbitrary driving torque is used. A waveform driving torque may be used.
　なお、本実施例では、それぞれのグループのパラメータ同定を、同一の矩形トルク用いて実施したが、これに限定するものでは無く、各グループの同定でその同定計算に適した駆動トルク(動作データ)を用いるなど、異なる駆動トルクを用いてもよい。 In this embodiment, the parameter identification of each group was performed using the same rectangular torque, but the present invention is not limited to this, and the driving torque (operation data) suitable for the identification calculation in the identification of each group. Different drive torques may be used, such as
　なお、本実施例では、評価値を実機とモデルのモータ回転速度応答とフライホイールの回転速度応答それぞれの差の二乗和の合計値としたが、これに限定するものでは無く、実機とモデルのモータ回転速度応答のみの差の二乗和であってもよいし、実機とモデルのモータ回転速度応答とフライホイールの回転速度応答あるいは、片方の回転速度をフーリエ変換し、結果得られるピーク振動数やその強度、あるいは平均値、あるいは実効値、あるいは単位時間あたりの変化量であってもよい。 In this embodiment, the evaluation value is the sum of squares of the difference between the motor rotation speed response of the actual machine and the model and the rotation speed response of the flywheel, but is not limited to this. It may be the sum of squares of the difference of only the motor speed response, or the Fourier speed of the motor speed response of the actual machine and the model and the speed response of the flywheel, or the rotation speed of one of them, and the resulting peak frequency or The intensity, the average value, the effective value, or the amount of change per unit time may be used.
　なお、本実施例では、制御器２の上流にパラメータ同定装置１を配置したが、図６に示すように制御器２と同じレベルであってもよいし、図７に示すようにモータアンプ３の内部に配置される構成でもよい。また、図８に示すように、パラメータ同定装置１をオフラインで配置し、予め記録してある指令を用いてパラメータ同定に必要な実機データを取得するための動作を実施し、その際のデータを記録しておき、パラメータ同定装置１はその記録されたデータを用いて、オフラインでパラメータ同定を実行する構成でもよい。 In the present embodiment, the parameter identification device 1 is arranged upstream of the controller 2, but it may be at the same level as the controller 2 as shown in FIG. 6, or the motor amplifier 3 as shown in FIG. The structure arrange | positioned inside may be sufficient. Further, as shown in FIG. 8, the parameter identification device 1 is arranged off-line, and an operation for acquiring actual machine data necessary for parameter identification is performed using a pre-recorded command. The parameter identification device 1 may be configured to perform parameter identification offline using the recorded data.
　なお、本実施例ではグループ１のパラメータ同定処理からグループ３のパラメータ同定処理を逐次的に実施したが、図９に示すようにグループ１のパラメータ同定処理からグループ３のパラメータ同定処理を並列に実行してパラメータ値の集結処理Ｆ１０６で、各グループで同定したパラメータ値を統合するパラメータ同定のフローであってもよいし、図１０に示すように逐次的なパラメータ同定の計算と並列なパラメータ同定計算を組み合わせて実行し、パラメータ値を統合するパラメータ同定のフローであってもよい。 In this embodiment, the group 1 parameter identification process is performed sequentially from the group 1 parameter identification process. However, as shown in FIG. 9, the group 1 parameter identification process is executed in parallel with the group 3 parameter identification process. Then, the parameter value gathering process F106 may be a parameter identification flow for integrating the parameter values identified in each group, or as shown in FIG. 10, the parameter identification calculation is performed in parallel with the sequential parameter identification calculation. May be executed in combination, and may be a parameter identification flow for integrating parameter values.
　図１１に従来の全パラメータを同時に同定する方法を用いた場合のパラメータ同定中のパラメータ値の推移の図の例を示す。 Fig. 11 shows an example of the transition of parameter values during parameter identification when the conventional method for identifying all parameters at the same time is used.
　図１２に従来のパラメータ同定値を用いたモータ回転速度応答の比較の図の例を示す。 Fig. 12 shows an example of a comparison of motor rotation speed responses using conventional parameter identification values.
　図１３に本発明の一実施形態によるパラメータ同定を用いた場合のパラメータ同定中のパラメータ値の推移の図の例を示す。 FIG. 13 shows an example of a transition diagram of parameter values during parameter identification when parameter identification according to an embodiment of the present invention is used.
　図１４に本発明の一実施形態によるパラメータ同定値を用いたモータ回転速度応答の比較の図の例を示す。 FIG. 14 shows an example of a comparison diagram of motor rotation speed responses using parameter identification values according to an embodiment of the present invention.
　図１１と図１３の比較から本発明の一実施形態によるパラメータ同定の場合は真値に近いパラメータの値が同定されていることがわかる。図１２と図１４の比較から従来手法によるパラメータ同定値を用いたモータ回転速度応答は実機とモデルとの差異がみられるが、本発明の一実施形態によるパラメータ同定値を用いたモータ回転速度応答は実機とモデルとの差異がほとんど無いことがわかる。 From the comparison between FIG. 11 and FIG. 13, in the case of parameter identification according to one embodiment of the present invention, it can be seen that a parameter value close to the true value is identified. From the comparison between FIG. 12 and FIG. 14, the motor rotation speed response using the parameter identification value according to the conventional method shows a difference between the actual machine and the model. Shows that there is almost no difference between the actual machine and the model.
　このようにして、パラメータを所定の特徴を持ってグループ分けし、グループ毎のパラメータ同定を実施し、直前のグループのパラメータ同定の結果を引き継ぎつつグループ毎のパラメータ同定を実施し、そのグループ毎の一連の計算を繰り返し実行することで、従来の全パラメータを一気に同定する場合に全パラメータの影響が平均化されることで発生していたパラメータどうしの応答への影響の干渉による同定精度悪化を防止することができ、計算モデルが多数の同定を必要とするパラメータを含み、これらが相互に計算結果へ干渉する場合であっても高精度な同定を実施できる。 In this way, parameters are grouped with predetermined characteristics, parameter identification is performed for each group, parameter identification for each group is performed while taking over the result of parameter identification of the immediately preceding group, and each group is identified. By repeatedly executing a series of calculations, when all parameters are identified at once, the effects of all parameters are averaged, preventing deterioration of identification accuracy due to interference with the effects of parameters on the response. Even if the calculation model includes a large number of parameters that require identification, and these interfere with each other in the calculation result, highly accurate identification can be performed.
　なお、本実施例ではパラメータを物理的な分類でグループ分けしたが、これに限定するものでは無く、パラメータの周波数感度を用いてグループ分けをおこなってもよい。図１５にこのグループ分けに用いるパラメータの周波数感度解析の例を示す。 In this embodiment, the parameters are grouped by physical classification. However, the present invention is not limited to this, and grouping may be performed using the frequency sensitivity of the parameters. FIG. 15 shows an example of frequency sensitivity analysis of parameters used for this grouping.
　図１５はモデルによるモータの回転速度をフーリエ変換行った結果をプロットしたものである。通常、すなわち各パラメータを初期値に設定した場合の応答は、周波数ｆ１にピークを持つ波形となっている。ここで、各パラメータを変化させて応答の変化を観測すると、Ｊ１、Ｊ２、ｃ１、ｃ２のパラメータを変化させた場合、周波数ｆ１のピークが変動する変化が観測される。一方、ｋｓ、ｃｓを変化させた場合、周波数ｆ１での変化は無く、周波数ｆ２でピークが現れる変化が観測される。これらの観測結果から、Ｊ１、Ｊ２、ｃ１、ｃ２が周波数ｆ１に感度があるパラメータであり、ｋｓ、ｃｓが周波数ｆ２に感度があるパラメータであることが分かり、それぞれ共通の応答に影響するパラメータであると思われる。このような周波数感度解析をもとにしてＪ１、Ｊ２、ｃ１、ｃ２からなる第一グループと、ｋｓ、ｃｓからなる第二グループにグループ分けをおこなってもよい。この手法により、パラメータのグループ分けを高速に実行することが可能となる。 FIG. 15 is a plot of the result of Fourier transform of the rotational speed of the motor according to the model. Usually, that is, when each parameter is set to an initial value, the response has a waveform having a peak at the frequency f1. Here, when the change of the response is observed by changing each parameter, when the parameters of J1, J2, c1, and c2 are changed, a change in which the peak of the frequency f1 fluctuates is observed. On the other hand, when ks and cs are changed, there is no change at the frequency f1, and a change in which a peak appears at the frequency f2 is observed. From these observation results, it can be seen that J1, J2, c1, and c2 are parameters that are sensitive to the frequency f1, and ks and cs are parameters that are sensitive to the frequency f2. It appears to be. Based on such frequency sensitivity analysis, a first group consisting of J1, J2, c1, and c2 and a second group consisting of ks and cs may be grouped. This technique makes it possible to execute parameter grouping at high speed.
　なお、本実施例ではパラメータを物理的な分類でグループ分けしたが、これに限定するものでは無く、同定するパラメータを知見データベースに基づいてグループ分けをおこなってもよい。 In this embodiment, the parameters are grouped by physical classification. However, the present invention is not limited to this, and the parameters to be identified may be grouped based on a knowledge database.
　図１６は、同定するパラメータを知見データベースに基づいてグループ分けする例である。換言すれば、制御対象の応答とパラメータを関連づけたデータベースに基づいてグループ分けしている。知見データベースＳ１６１には例えば、知見１として「トルク印加時の速度応答の差異は、イナーシャと摩擦損失によるものである」。知見２として「トルク急変時に速度波形に生じる振動はシャフトのねじり特性によるものである」。というような知見を登録しておく。知見１および知見２により、知見データベースＳ１６１がＪ１、Ｊ２、ｃ１、ｃ２からなる第一グループＳ１６２およびｋｓ、ｃｓからなる第二グループＳ１６３にグループ分けされる。 FIG. 16 is an example of grouping parameters to be identified based on a knowledge database. In other words, grouping is performed based on a database that associates responses to be controlled with parameters. For example, in the knowledge database S161, as knowledge 1, “the difference in speed response at the time of torque application is due to inertia and friction loss”. According to knowledge 2, “vibration that occurs in the speed waveform during sudden torque changes is due to the torsional characteristics of the shaft” Such knowledge is registered. Based on knowledge 1 and knowledge 2, knowledge database S161 is grouped into first group S162 consisting of J1, J2, c1, and c2 and second group S163 consisting of ks and cs.
　図１７にこれらの知見と速度応答の関係を示す。パラメータのグループ分けを行う際に、この知見データベースを参照して、Ｊ１、Ｊ２、ｃ１、ｃ２からなる第一グループと、ｋｓ、ｃｓからなる第二グループにグループ分けをおこなってもよい。この手法により、パラメータのグループ分けを過去の知見に基づいて的確に実行することが可能となる。 Fig. 17 shows the relationship between these findings and the speed response. When performing parameter grouping, referring to this knowledge database, grouping may be performed into a first group consisting of J1, J2, c1, and c2 and a second group consisting of ks and cs. This technique makes it possible to accurately execute parameter grouping based on past knowledge.
　なお、本実施例では、制御器２の上流や同じレベルにパラメータ同定装置１を配置したが、図１８に示すようにモータアンプ３と同じレベルであっても良い。この場合、例えば、本発明のパラメータ同定装置はモータアンプに接続するパーソナルコンピュータとパーソナルコンピュータ上で動作するソフトウェアの形態をとることができる。パーソナルコンピュータとモータアンプを通信回線で接続し、パーソナルコンピュータからモータアンプに指令を出したり、モータアンプ内の情報をパーソナルコンピュータで取得するようにし、本実施例で説明している同定に用いる一連の制御モードの切り換えやパラメータの同定処理、また、データの記録等をパーソナルコンピュータとモータアンプを連携させて実施できるようにしても良い。また、パーソナルコンピュータ上の様々の取得データや同定結果などを、パーソナルコンピュータのまた別の通信回線を通じて他のシステムへ持ち出したり、パーソナルコンピュータ自体を持ち運ぶ事でデータを他のシステムへ持ち出し、その、他のシステムで、さらに本発明の同定計算を、例えば、さらに高精度に実行して、より高精度なパラメータを求めても良い。 In this embodiment, the parameter identification device 1 is arranged upstream of the controller 2 or at the same level, but it may be at the same level as the motor amplifier 3 as shown in FIG. In this case, for example, the parameter identification device of the present invention can take the form of a personal computer connected to the motor amplifier and software operating on the personal computer. A personal computer and a motor amplifier are connected via a communication line, a command is sent from the personal computer to the motor amplifier, and information in the motor amplifier is acquired by the personal computer. Control mode switching, parameter identification processing, data recording, and the like may be performed in cooperation with a personal computer and a motor amplifier. Also, various acquired data and identification results on a personal computer can be taken to another system through another communication line of the personal computer, or the data can be taken to another system by carrying the personal computer itself. In this system, the identification calculation of the present invention may be further performed, for example, with higher accuracy to obtain parameters with higher accuracy.
　なお、本実施例では、制御器２の上流や同じレベルにパラメータ同定装置１を配置したが、図１９に示すようにモータアンプとクラウドで接続される形態であっても良い、この場合、例えば、パラメータ同定は、管理部門に設置されたパーソナルコンピュータなどで実施することができる。これにより、物理的な制約に左右されずにパラメータ同定装置を設置することができる。また、クラウド利用により複数の装置を対象として、統括的に管理部門のパーソナルコンピュータでパラメータ同定を実施することができる。また、パラメータ同定装置を搭載したパーソナルコンピュータ上で、同定結果を監視することにより、装置の状態の変化を感知すること、複数の装置から成るシステム全体の監視を実行することができる。 In this embodiment, the parameter identification device 1 is arranged upstream of the controller 2 or at the same level. However, as shown in FIG. 19, it may be connected to a motor amplifier via a cloud. In this case, for example, The parameter identification can be performed by a personal computer or the like installed in the management department. Thereby, a parameter identification device can be installed without being influenced by physical restrictions. In addition, parameter identification can be performed on a personal computer in the management department in an integrated manner for a plurality of devices by using the cloud. Further, by monitoring the identification result on a personal computer equipped with a parameter identification device, it is possible to sense a change in the state of the device and to monitor the entire system composed of a plurality of devices.
　以上、本発明によれば、計算機上でモデルを用いた計算により対象のふるまいを把握して機器の設計を改善するモデルベース開発技術に関し、特にモデルのパラメータ同定の方法およびその装置に関して、計算モデルが多数の同定を必要とするパラメータを含み、これらが相互に計算結果へ干渉する場合であっても高精度なパラメータ同定を実現できる手法および装置を提供することができる。 As described above, according to the present invention, the present invention relates to a model-based development technique for improving the design of a device by grasping the behavior of a target by calculation using a model on a computer, and in particular, with respect to a method and apparatus for identifying a parameter of a model. Can include a number of parameters that require identification, and can provide a method and apparatus that can realize highly accurate parameter identification even when these parameters interfere with each other in the calculation results.
１　パラメータ同定装置
２　制御器
３　モータアンプ
４　モータ
５　機械
３１　第一の回転イナーシャ
３３　第二の回転イナーシャ
３４　第一の粘性減衰
３５　第二の粘性減衰
３２１　シャフトの剛性
３２２　シャフトの粘性減衰
Ｓ１　実測データ記憶部
Ｓ２　パラメータ変数記憶部
Ｓ３　パラメータグループ分け部
Ｓ４　パラメータ同定部
Ｓ５　パラメータ値記憶部
Ｓ４１　モデル実行部
Ｓ４２　同定実行部 DESCRIPTION OF SYMBOLS 1 Parameter identification apparatus 2 Controller 3 Motor amplifier 4 Motor 5 Machine 31 1st rotation inertia 33 2nd rotation inertia 34 1st viscosity damping 35 2nd viscosity damping 321 Shaft rigidity 322 Shaft viscosity damping S1 Actual measurement data Storage unit S2 Parameter variable storage unit S3 Parameter grouping unit S4 Parameter identification unit S5 Parameter value storage unit S41 Model execution unit S42 Identification execution unit

Claims (6)

  1. 　制御対象の挙動を計算機上で予想するために用いる計算モデルの複数のパラメータを同定する方法であって、
    　同定する前記複数のパラメータを複数のグループにグループ分けする工程と、
    　前記複数のグループ毎に前記複数のグループ内のパラメータを同定する工程と、
    　前記複数のグループ内のあるグループでパラメータが同定されることにより得られたパラメータ値を、前記あるグループ中のパラメータが同定された以降にパラメータを同定するグループの計算に引き継いで利用しながら、前記複数のグループ毎に前記複数のグループ内のパラメータの同定を繰返す工程と、
    　所定の終了条件になったところで前記繰返しを終了し、前記繰返しにより得られたパラメータ値を同定結果とする工程と、を有する
    　複数のパラメータを同定する方法。     A method for identifying a plurality of parameters of a calculation model used to predict the behavior of a controlled object on a computer,
    Grouping the plurality of parameters to be identified into a plurality of groups;
    Identifying a parameter in the plurality of groups for each of the plurality of groups;
    The parameter value obtained by identifying a parameter in a certain group of the plurality of groups is used by taking over the calculation of the group for identifying the parameter after the parameter in the certain group is identified, Repeating the identification of parameters in the plurality of groups for each of a plurality of groups;
    A method of identifying a plurality of parameters, comprising: ending the repetition when a predetermined end condition is satisfied, and using the parameter value obtained by the repetition as an identification result.
  2. 　請求項１において、
    　前記複数のパラメータの周波数感度を用いて、同定する前記複数のパラメータを複数のグループにグループ分けする
    　複数のパラメータを同定する方法。     In claim 1,
    A method for identifying a plurality of parameters, wherein the plurality of parameters to be identified are grouped into a plurality of groups using frequency sensitivity of the plurality of parameters.
  3. 　請求項１において、
    　前記制御対象の応答とパラメータを関連づけたデータベースに基づいて、同定する前記複数のパラメータを複数のグループにグループ分けする
    　複数のパラメータを同定する方法。     In claim 1,
    A method for identifying a plurality of parameters, wherein the plurality of parameters to be identified are grouped into a plurality of groups based on a database in which the response to be controlled and the parameters are associated with each other.
  4. 　請求項１において、
    　前記複数のグループ毎に前記複数のグループ内のパラメータを同定する際、前記パラメータの同定に適した動作データを用いる
    　複数のパラメータを同定する方法。     In claim 1,
    A method of identifying a plurality of parameters using operation data suitable for identifying the parameters when identifying parameters in the plurality of groups for each of the plurality of groups.
  5. 　請求項１において、
    　前記計算モデルの対象が電気モータで駆動される機械装置であって、
    　前記複数のグループ毎に前記複数のグループ内のパラメータを同定する際に、前記電気モータの回転速度または回転角度を用いる
    　複数のパラメータを同定する方法。     In claim 1,
    The object of the calculation model is a mechanical device driven by an electric motor,
    A method of identifying a plurality of parameters using a rotation speed or a rotation angle of the electric motor when identifying parameters in the plurality of groups for each of the plurality of groups.
  6. 　請求項１乃至５のいずれかの複数のパラメータを同定する方法を用いた複数のパラメータ同定装置。 A plurality of parameter identification devices using the method for identifying a plurality of parameters according to claim 1.
PCT/JP2015/072988 2014-09-19 2015-08-17 Method for identifying plurality of parameters and multiple parameters identification device using method for identifying plurality of parameters WO2016042958A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2014-190661 2014-09-19
JP2014190661A JP2017207792A (en) 2014-09-19 2014-09-19 Method for identifying a plurality of parameters and a plurality of parameter identifying devices using the same

Publications (1)

Publication Number Publication Date
WO2016042958A1 true WO2016042958A1 (en) 2016-03-24

Family

ID=55533000

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2015/072988 WO2016042958A1 (en) 2014-09-19 2015-08-17 Method for identifying plurality of parameters and multiple parameters identification device using method for identifying plurality of parameters

Country Status (2)

Country Link
JP (1) JP2017207792A (en)
WO (1) WO2016042958A1 (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH07152429A (en) * 1993-11-29 1995-06-16 Mitsubishi Heavy Ind Ltd Parameter identifying device
JP2004276125A (en) * 2003-03-12 2004-10-07 Mitsubishi Electric Corp Load parameter identification method
JP2008228360A (en) * 2007-03-08 2008-09-25 Hitachi Industrial Equipment Systems Co Ltd Motor controller and motor control system

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH07152429A (en) * 1993-11-29 1995-06-16 Mitsubishi Heavy Ind Ltd Parameter identifying device
JP2004276125A (en) * 2003-03-12 2004-10-07 Mitsubishi Electric Corp Load parameter identification method
JP2008228360A (en) * 2007-03-08 2008-09-25 Hitachi Industrial Equipment Systems Co Ltd Motor controller and motor control system

Also Published As

Publication number Publication date
JP2017207792A (en) 2017-11-24

Similar Documents

Publication Publication Date Title
CN112798280A (en) Rolling bearing fault diagnosis method and system
CN109219738A (en) Apparatus for diagnosis of abnormality and abnormality diagnostic method
JP6324641B1 (en) Abnormality diagnosis apparatus and abnormality diagnosis method
EP3760999A1 (en) Abnormality type determining device and abnormality type determining method
EP3764183A1 (en) Managing health condition of a rotating system
JP6850458B1 (en) AI-equipped motor state estimation system and machine learning method for motor models
CN114260892A (en) Elastic joint torque control method and device, readable storage medium and robot
WO2016042958A1 (en) Method for identifying plurality of parameters and multiple parameters identification device using method for identifying plurality of parameters
Jouhet et al. Model-based fault detection of permanent magnet synchronous motors of drones using current sensors
Haikonen et al. Small-scale test bench of maritime thruster for digital twin research
Huang et al. Sensorless vibration harmonic estimation of servo system based on the disturbance torque observer
JP4019710B2 (en) Engine bench system
US10216151B2 (en) Power consumption-amount estimation apparatus
CN109844658B (en) Parameterized automatic optimization of motion control device
CN110262576A (en) A kind of rack control method and device
JPWO2020039661A1 (en) Abnormality diagnostic device
CN113806876B (en) Robot state judging method and device
EP3135953A1 (en) Active vibration damping device and design method
Aditiya et al. Fault diagnosis system of rotating machines using Hidden Markov Model (HMM)
CN110208691B (en) Motor load parameter detection method and device
JP2022023582A (en) Servo motor device and control method
KR20170050191A (en) Multi-axis dynamometer system and control method of the same
Kuperman Testing motion controllers robustness by emulating electrical and mechanical parameter variations of motor drives
JP7281629B2 (en) motor drive
JP7257940B2 (en) Control device

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15841271

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15841271

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: JP