Multi-Model Control System Tuning via Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Designers of control systems face challenges in constructing and analyzing models that incorporate multiple constraints, including both flexible and inflexible ones, which complicates the optimization of control system performance.
Innovation Solution
A method is described where a control system model is created with tunable parameters and constraints, allowing users to input both hard and soft constraints. An optimization algorithm is applied to determine parameter values that maximize performance subject to these constraints, using a technical computing environment to simulate and adjust the model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple constraints (both flexible and inflexible) are incorporated into the control system model, then the system can satisfy more design requirements, but the complexity of constructing and analyzing the model increases
Solution Approach 1:
The patent segments the control system model into multiple independent models, each representing different operating conditions or constraints. This allows the complex multi-constraint problem to be divided into manageable sub-problems that can be analyzed separately while still satisfying all design requirements collectively.
Solution Approach 2:
The patent utilizes parameter changes by adjusting tunable parameters across different models to satisfy varying constraints. By changing parameters such as controller gains, filter coefficients, or operating point values in each model, the system can adapt to different design requirements without increasing overall structural complexity.
2Reliability
If an optimization algorithm is applied to determine parameter values that maximize performance subject to constraints, then the control system performance is improved, but the computational time and resources required increase
Solution Approach 1:
The patent applies preliminary action by pre-defining multiple models with different operating conditions and constraints before optimization. This preparation work is done upfront, so that when optimization is performed, the algorithm only needs to adjust parameters within pre-established boundaries rather than exploring the entire parameter space, reducing computational time.
Solution Approach 2:
The patent employs dynamics by using adaptive optimization that adjusts the optimization process itself based on problem characteristics. The optimization algorithm dynamically selects appropriate methods and parameters based on the specific constraints and performance goals, improving efficiency while maintaining solution quality.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
ABSTRACT A device receives a control system model that includes a fixed portion that models elements of a control system to be controlled and a tunable portion that models elements of the control system used to control the elements modeled by the fixed portion. The device receives information that identifies a tunable parameter of the tunable portion of the control system model, a hard constraint associated with the control system model, and a soft constraint associated with the control system model. The hard constraint identifies a first constraint that is to be satisfied, and the soft constraint identifies a second constraint that is to be reduced. The device calculates a parameter value for the tunable parameter by applying an optimization algorithm to the control system model, based on the control system model, the tunable parameter, the hard constraint, and the soft constraint. The device provides the parameter value.