Machine model predicting device of motor control device

• US 20040245959A1
• Filed: 03/03/2004
• Published: 12/09/2004
• Est. Priority Date: 09/04/2001
• Status: Active Grant

• Current Assignee: Kabushiki Kaisha Yaskawa Denki
• Original Assignee: Kabushiki Kaisha Yaskawa Denki
• Inventors: Komiya, Takehiko
• Granted Patent: US 6,903,528 B2
• Time in Patent Office: Days
• US Class Current: 318/607
• CPC Class Codes: G05B 2219/34048 Fourier transformation, analysis, fft
• CPC Class Codes: H02P 21/13 Observer control, e.g. using Luenberger observers or Kalman filters

Abstract

In an electric motor control apparatus including an electric motor 4 for driving a load machine, a rotation detector 3 for detecting a rotating angle of the electric motor 4, and a servo control device 2 for controlling the electric motor 4, a calculating device 1 automatically calculates protruded shapes to be a resonance frequency and an anti-resonance frequency from frequency characteristics obtained from an operation command signal 8 and a rotation detector signal 9, and furthermore, calculates errors of frequency characteristics calculated by frequency characteristic equations 20 for a 2-inertia model and a rigid body model from a frequency characteristic obtained by a measurement, and compares the minimum values of calculated values in the frequency characteristics of the 2-inertia model and the rigid body model and measured values respectively, thereby automatically modeling the characteristic of the machine.

Consequently, it is possible to provide a machine model estimating device of an electric motor control apparatus which can easily estimate a machine model without using an expensive measuring apparatus even if an operator has neither advanced expertise nor experiences, and furthermore, is inexpensive.

Claims

1. A machine model estimating device of an electric motor control apparatus comprising an electric motor for driving a load machine, a rotation detector for detecting a rotating angle of the electric motor, and a servo control device for controlling the electric motor, comprising

a calculating device for outputting an operation command signal for operating the electric motor to the servo control device, and frequency characteristic equations for a rigid body model and an N-inertia model (N is an integer which is equal to or greater than 2) which are previously input to the calculating device,

wherein the calculating device includes a frequency characteristic measuring section for measuring a frequency characteristic from the operation command signal and a signal of the rotation detector input from the servo control device to the calculating device, a frequency characteristic peak detecting section for automatically calculating protruded shapes to be a resonance frequency and an anti-resonance frequency from a shape of the frequency characteristic measured by the frequency characteristic measuring section, an attenuation estimation value analyzing section for estimating an attenuation from the resonance frequency and the anti-resonance frequency which are detected by the frequency characteristic peak detecting section, a frequency characteristic error calculating section for calculating errors of the frequency characteristics calculated in the frequency characteristic equation for the N-inertia model and the frequency characteristic equation for the rigid body model from the frequency characteristic obtained by the measurement respectively, and a machine model deciding section for comparing a minimum error of a calculated value of the frequency characteristic of the N-inertia model which is obtained in the frequency characteristic error calculating section and a measured value with a minimum error of a calculated value of the frequency characteristic of the rigid body model and a measured value and deciding, as an actual model, either of the models which has a smaller error.

2. The machine model estimating device of an electric motor control apparatus according to claim 1, wherein the frequency characteristic error calculating section carries out curve fitting of the frequency characteristic obtained from the operation command signal and the signal of the rotation detector to the frequency characteristic equation, thereby calculating an error of the calculated value of the frequency characteristic and the measured value.

3. A machine model estimating device of an electric motor control apparatus comprising an electric motor for driving a load machine, a vibration detector for detecting a vibration of the load machine, and a servo control device for controlling the electric motor, comprising

a calculating device for outputting an operation command signal for operating the electric motor to the servo control device, and frequency characteristic equations of a rigid body model and an N-inertia model (N is an integer which is equal to or greater than 2) which are previously input to the calculating device,

wherein the calculating device includes a frequency characteristic measuring section for measuring a frequency characteristic from the operation command signal and a signal of the vibration detector input from the servo control device to the calculating device, a frequency characteristic peak detecting section for automatically calculating protruded shapes to be a resonance frequency and an anti-resonance frequency from a shape of the frequency characteristic measured by the frequency characteristic measuring section, an attenuation estimation value analyzing section for estimating an attenuation from the resonance frequency and the anti-resonance frequency which are detected by the frequency characteristic peak detecting section, a frequency characteristic error calculating section for calculating errors of the frequency characteristics calculated in the frequency characteristic equation for the N-inertia model and the frequency characteristic equation for the rigid body model from the frequency characteristic obtained by the measurement respectively, and a machine model deciding section for comparing a minimum error of a calculated value of the frequency characteristic of the N-inertia model which is obtained in the frequency characteristic error calculating section and a measured value with a minimum error of a calculated value of the frequency characteristic of the rigid body model and a measured value and deciding, as an actual model, either of the models which has a smaller error.

4. The machine model estimating device of an electric motor control apparatus according to claim 3, wherein the frequency characteristic error calculating section carries out curve fitting of the frequency characteristic obtained from the operation command signal and the signal of the vibration detector to the frequency characteristic equation, thereby calculating an error of the calculated value of the frequency characteristic and the measured value.