Powertrain Model-Predictive Control Using Preview Information
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The hybrid control problem in powertrain systems, where continuous actuators transition to or from discrete modes, poses challenges in accurately and efficiently calibrating robust control solutions due to the vast number of control variables, leading to inefficiencies in torque delivery and energy management.
Innovation Solution
A model-predictive control (MPC) approach is implemented using a controller that incorporates preview information from multiple sensors to anticipate future torque demands, allowing proactive control actions and optimizing control over a future window, thereby improving the efficiency and smoothness of powertrain system operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If pre-programmed lookup tables or calibrated mode transition schedules are used to control hybrid powertrain systems, then the control implementation is simple, but the accuracy and robustness of control solutions deteriorates due to the vast number of control variables
Solution Approach 1:
The patent transforms the control approach from static lookup tables to dynamic model-predictive control by changing the parameters from fixed pre-programmed values to real-time optimized control variables. The MPC controller continuously adjusts control inputs based on current system state and future torque demand predictions, enabling accurate control despite the vast number of control variables in hybrid powertrain systems.
2Speed
If control actions are based on instantaneous torque or speed demand, then the response is immediate, but the efficiency and smoothness of mode transitions deteriorates due to lack of anticipatory control
Solution Approach 1:
The patent implements preliminary action by using model-predictive control to anticipate future torque demands and prepare for mode transitions in advance. The controller predicts future operating conditions and proactively adjusts control inputs to optimize mode transitions, thereby improving transition efficiency and smoothness while maintaining immediate responsiveness to actual torque demands.
3Ease of operation
If discrete mode transitions are implemented without predictive control, then the system operation is straightforward, but the noise and vibration during transitions increases due to non-optimal transition timing
Solution Approach 1:
The patent employs feedback mechanisms within the model-predictive control framework to continuously monitor system state and adjust control inputs for optimal mode transitions. The controller uses real-time feedback on system operating conditions to determine the optimal timing for discrete mode transitions, thereby minimizing torque transients and reducing noise and vibration while maintaining straightforward system operation.
Data Source
AI summary
A method for controlling continuous and discrete actuators (e.g., modes) in a powertrain system includes receiving preview information from a sensor(s) describing an upcoming dynamic state at a future time point, and providing control inputs for the actuators to a controller that includes the preview information. The input set collectively describes a future torque or speed output state at the future time point. The controller processes the input set via a dynamical predictive model, in real time, to determine control solutions to take at the present time point for implementing the dynamic state at the future time point. A lowest opportunity cost control solution is determined and optimized. The controller executes the optimized solution at the present time step.