Engine Airpath MPC Recalibration Without Model Rebuild
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The complexity of engine airpath control systems, with increasing features leading to non-linear control problems and challenges in recalibrating model predictive control (MPC) solutions, results in significant time and resource consumption for calibration changes.
Innovation Solution
The approach involves modifying the calibration of the MPC calculation by updating the coefficients of the cost function, particularly the Hessian matrix and linear part, without rebuilding the model, allowing for efficient recalibration and reducing computational complexity by using a fixed problem structure and quadratic programming.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If switched linear MPC problems are used to reduce computational complexity, then computation time is reduced, but recalibration time increases significantly when calibration changes are needed
Solution Approach 1:
The patent segments the MPC problem into a fixed structural framework (the switched linear MPC formulation) and variable calibration parameters (cost function weights and constraints). This allows the computational framework to be reused while only the parameter values need recalibration, significantly reducing recalibration time when system requirements change.
Solution Approach 2:
The patent enables efficient recalibration by allowing direct modification of calibration parameters (cost function weights, constraints) without changing the underlying MPC mathematical structure. This parameter-based recalibration approach is much faster than complete model rebuilding while adapting to new system requirements.
2Reliability
If full MPC is used for nonlinear control solutions, then control performance is improved, but computational complexity increases significantly
Solution Approach 1:
The patent segments the complex nonlinear MPC problem into multiple simpler linear MPC sub-problems, each valid in a localized operating region. This segmentation maintains control performance by ensuring appropriate linear models are active in their valid ranges while reducing computational burden through simpler linear optimization.
Solution Approach 2:
The patent implements a dynamic switching mechanism that selects among multiple linear MPC controllers based on current operating conditions. This dynamic approach maintains high control performance by adapting to changing system states while keeping individual controller computations simple and fast.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods and systems for controlling a system such as an engine having an airflow system. A model predictive control calculation is configured in an off-line mode, having a linear part and a non-linear part. In an on-line mode, the linear part of the MPC and/or a Hessian matrix used with the MPC is modified responsive to special modes or other operating changes or conditions. The online mode is configured to respond to changing modes or conditions without requiring recalculation of the MPC. Certain changes of conditions and modes are used to modify feedforward, while others modify responsiveness.