Internal Combustion Engine Control With Runtime Model Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for controlling internal combustion engines require significant tuning effort due to the large number of characteristic curves and maps, and there is a lack of effective methods for adapting data-based functional models at runtime to account for manufacturing variations and component deviations.
Innovation Solution
The method involves calculating injection and gas path setpoints using combustion and gas path models, respectively, and adapting these models during engine operation using Gaussian process models to minimize a quality measure within a prediction horizon, incorporating data from both normal and extreme operating conditions, with confidence interval assessment and data point management to ensure accuracy and reduce computation time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If characteristic curves and maps are used for engine control, then the engine operation can be controlled according to desired power output, but the tuning effort increases significantly due to the large number of characteristic curves and maps
Solution Approach 1:
The patent replaces the traditional mechanical approach of using numerous characteristic curves and maps with a data-based functional model (neural network) that can be adapted at runtime. This substitution reduces the manual tuning effort while maintaining the ability to control engine operation according to desired power output.
Solution Approach 2:
The patent enables runtime adaptation of model parameters in the data-based functional model, allowing the system to adjust to manufacturing variations and component deviations without requiring extensive pre-tuning of characteristic curves. This dynamic parameter adjustment reduces the initial tuning effort while maintaining control accuracy.
2Stability of the object's composition
If model parameters are preset and not changed after vehicle commissioning, then the model remains stable, but it cannot account for manufacturing variations, component deviations, or changes in system behavior over time
Solution Approach 1:
The patent transforms the static model parameters into dynamic parameters that can be adapted at runtime based on actual engine behavior and measured data. This allows the system to maintain stability through continuous adaptation rather than fixed parameters, accommodating manufacturing variations and component deviations while preserving overall system stability.
Solution Approach 2:
The patent implements a feedback mechanism where the data-based functional model is continuously adapted using measured engine data and performance deviations. This feedback loop enables the system to account for manufacturing variations, component deviations, and aging effects while maintaining stable and accurate engine control over the vehicle's lifetime.
3Adaptability or versatility
If online adaptation of model parameters is performed during operation, then manufacturing variations and component deviations can be compensated, but computation time and processing load increase
Solution Approach 1:
The patent performs partial adaptation by focusing computational resources on adapting only the critical model parameters that have the greatest impact on engine performance and emission control. This selective adaptation approach reduces computation time while still effectively compensating for manufacturing variations and component deviations.
4Device complexity
If data-based functional models are used instead of characteristic curves, then tuning effort is reduced, but the models require adaptation to account for manufacturing variations and component deviations
Solution Approach 1:
The patent implements self-service adaptation where the data-based functional model automatically adjusts its parameters using measured engine data and performance feedback. This automated self-adaptation process reduces the need for manual tuning while effectively accounting for manufacturing variations, component deviations, and system aging, balancing reduced complexity with appropriate automation.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method for the model-based control and regulation of an internal combustion engine (1), in which, as a function of a set torque (M(SOLL)), injection system set values for controlling the injection system actuators are calculated via a combustion model (20), and gas path set values for controlling the gas path actuators are calculated via a gas path model (22), the combustion model (20) being adapted during operation of the internal combustion engine (1). A measure of quality is calculated by an optimizer (23) as a function of the injection system set values and the gas path set values. The measure of quality is minimized by the optimizer (23) by changing the injection system set values and gas path set values within a prediction horizon, and the injection system set values and gas path set values are set by the optimizer (23) as critical for adjusting the operating point of the internal combustion engine (1) by using the minimized measure of quality.