Model Predictive Control Using Periodic Error Amplitude Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional model predictive control devices require a complex model of external disturbances, making it difficult to apply them to a wide range of machines and devices, especially when the amplitude of the waveform representing changes in predicted error is not constant over time.
Innovation Solution
A model predictive control device that determines the presence of periodicity in predicted errors and derives an equation to correct the predicted value using the error from previous periods, without requiring a model of external disturbances, by calculating amplitude variation ratios and using coefficients to represent changes in amplitude over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a model of external disturbances is prepared in advance, then control accuracy can be improved, but device complexity and difficulty of application increase significantly
Solution Approach 1:
The patent uses historical predicted error data from previous periods as a copy of the disturbance pattern, replacing the need for complex external disturbance models. By copying and adapting past error patterns, the system achieves disturbance compensation without requiring sophisticated modeling of external disturbance sources.
Solution Approach 2:
The system uses its own historical predicted error data to compensate for current disturbances, making the system self-sufficient. Instead of requiring external disturbance models or additional sensors, the controller leverages its own past performance data to improve current control accuracy.
2Reliability
If conventional model predictive control is used, then control prediction can be achieved, but it fails when amplitude of predicted error waveform is not constant
Solution Approach 1:
The patent introduces dynamic amplitude correction by calculating amplitude variation ratios between different periods. Instead of assuming constant amplitude, the system dynamically adjusts the predicted error based on observed amplitude changes, making the control reliable even when disturbance amplitudes vary over time.
Solution Approach 2:
The system changes the parameter representation of predicted error from a static model to a dynamic one by incorporating amplitude variation ratios. This parameter transformation allows the system to adapt to varying disturbance amplitudes while maintaining control prediction reliability.
3Ease of manufacture
If amplitude variation is not considered, then control calculation is simpler, but control accuracy deteriorates when periodic errors have varying amplitude
Solution Approach 1:
The patent segments the predicted error correction into distinct components: base predicted error from the model and amplitude correction based on historical variation ratios. This segmentation allows the system to maintain computational simplicity while adding accuracy through modular error correction components.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Compensation for an external disturbance can be applied without requiring a model of the external disturbance, and, particularly, the amount of control can be accurately predicted even when the amplitude of a waveform representing changes in a predicted error with respect to time is not constant with respect to time in a case in which the predicted error that is an error between an actually-measured value of the amount of control of a control target and a predicted value periodically changes. A controller (200) compensates a predicted value of the current amount of control using a predicted error of a previous period by taking changes in the amplitude of a waveform representing a predicted error that is an error between an actually-measured value of the amount of control and a predicted value into account.