Servo Control Gain Optimization via Online Frequency Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional servo control systems lack the ability to quantitatively evaluate the difference from a reference transfer function directly in the frequency domain online, which is essential for optimizing control gain and achieving stability and high-speed response.
Innovation Solution
A servo control apparatus with a speed control loop that includes a sinusoidal disturbance input unit, frequency characteristic calculation unit, reference characteristic generation unit, and gain adjusting unit, which applies sinusoidal sweeps to estimate gain and phase, calculates an evaluation function value based on measured and reference frequency characteristics, and adjusts the speed control loop gain online to achieve desired evaluation function values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional reference model fitting methods are used to optimize control gain, then control parameters can be adjusted, but the ability to quantitatively evaluate frequency characteristics directly in the frequency domain online is lost
Solution Approach 1:
The patent implements an online feedback mechanism where frequency characteristics are continuously measured during operation, compared against reference characteristics, and used to automatically adjust control parameters. The evaluation function calculates the difference between actual and reference frequency characteristics, feeding this information back to the parameter adjustment unit for real-time optimization without requiring complex offline modeling.
Solution Approach 2:
The patent replaces conventional time-domain reference model fitting methods with direct frequency-domain measurement and evaluation. Instead of using complex mathematical fitting procedures in the time domain, the system directly measures frequency characteristics using sinusoidal inputs and evaluates them in the frequency domain, simplifying the overall control system architecture.
2Productivity
If time-domain reference model fitting is used for control optimization, then parameter adjustment is possible, but direct online evaluation of frequency characteristics is not achieved
Solution Approach 1:
The patent pre-establishes reference frequency characteristics that represent ideal control behavior. These reference characteristics are stored and used as the basis for online evaluation, allowing the system to quickly compare actual performance against known good performance without requiring complex real-time optimization calculations, thus improving both speed and reliability.
Solution Approach 2:
The patent changes the domain of evaluation from time-domain fitting to frequency-domain measurement. By measuring frequency characteristics directly using sinusoidal inputs and evaluating them in the frequency domain, the system achieves faster online optimization while maintaining stability through direct comparison with reference frequency characteristics rather than indirect time-domain fitting.
3Measurement precision
If conventional frequency measurement methods are used, then frequency characteristics can be obtained, but online optimization of control gain using evaluation functions is not possible
Solution Approach 1:
The patent implements a self-service optimization system where the control system automatically measures its own frequency characteristics, evaluates them against reference characteristics using an evaluation function, and adjusts its own parameters without external intervention. This automated self-optimization process maintains high measurement accuracy while making parameter adjustment extremely easy, as the system performs all optimizations autonomously.
Data Source
AI summary
A servo control apparatus includes a speed command generation unit, a torque command generation unit, a speed detection unit, a speed control loop, a speed control loop gain setting unit, a sinusoidal disturbance input unit which applies a sinusoidal sweep to the speed control loop, a frequency characteristic calculation unit which estimates the gain and phase of a speed control loop input/output signal from output of the speed control loop when a sine wave is input to the speed control loop, a reference characteristic generation unit which generates frequency characteristics for a feed axis, and a gain adjusting unit which adjusts the speed control loop gain. The gain adjusting unit uses the reference frequency characteristics and actually measured frequency characteristics to calculate an evaluation function value for each frequency sweep, and adjusts online the speed control loop gain to achieve a desired evaluation function value.