Servo Control Switching Between Sensor and Model for Vibration Damping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing servo control devices face challenges in damping vibration effectively, especially when the controlled object's position is measured only within a portion of its movability range, leading to delayed vibration damping and inaccurate tracking of target trajectories.
Innovation Solution
A control device that employs model predictive control using a dynamics model to generate manipulated variables, switching between measured values and model output values based on the controlled object's position range, and includes a third mode to smooth torque variations during mode transitions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If vibration detection is performed using torque command value oscillation, then vibration damping control can be initiated, but the control start is delayed because vibration must be detected after the object stops moving
Solution Approach 1:
The system performs preliminary vibration detection during the movement phase using the measurement sensor before stopping. The vibration characteristics are detected and stored in advance, so that vibration damping control can be initiated immediately upon stopping without waiting for post-stop vibration detection, thereby eliminating the control start delay while maintaining damping effectiveness
2Device complexity
If position measurement is limited to a portion of the movability range, then the measurement sensor can be simplified, but vibration damping accuracy deteriorates when the controlled object is outside the measurement range
Solution Approach 1:
The control range is segmented into two zones: a first control range where the measurement sensor operates and provides accurate position feedback, and a second control range where the measurement sensor cannot measure. The controller switches between different control algorithms based on the current position zone, enabling accurate vibration damping in both regions while maintaining a simple measurement sensor configuration
Solution Approach 2:
A model predictive control algorithm acts as an intermediary for positions outside the measurement range. The model predictive control uses a dynamics model to predict and compensate for vibrations when the controlled object is in the second control range, bridging the gap caused by the measurement sensor's limited range and maintaining positioning accuracy without requiring sensor modification
3Manufacturing precision
If model predictive control switches between using measured values and model output values, then control accuracy is maintained across the full movability range, but torque fluctuations occur during mode transitions
Solution Approach 1:
Before switching between control modes, the system performs preliminary calculations to predict the transition timing and prepares compensation values in advance. The controller calculates the difference between the two control algorithms' output torque values beforehand and applies smoothing compensation during the transition period, cushioning the torque fluctuations before they fully manifest and maintaining trajectory tracking accuracy throughout the mode switch
Data Source
AI summary
A control device is connected to a servo mechanism that drives a controlled object and outputs a manipulated variable to the servo mechanism so that a controlled variable tracks a target trajectory. The control device includes a controller and a sensor. The controller acquires a measured value from the sensor and performs model predictive control for each control period using a dynamics model representing a relationship between the manipulated variable and the position of the controlled object to generate the manipulated variable to be output to the servo mechanism. The sensor measures the position of the controlled object. The controller performs model predictive control in a first mode using the measured value when the controlled object has a position within the range, and performs model predictive control in a second mode using an output value of the dynamics model when the controlled object has a position outside the range.