Air-Fuel Ratio Learning Control for Engine Convergence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing air-fuel ratio control systems for engines face inefficiencies in learning air-fuel ratio values across multiple operating regions, leading to prolonged time requirements for achieving optimal air-fuel ratio convergence, especially after initial learning is cleared.
Innovation Solution
An air-fuel ratio control apparatus that includes an air-fuel ratio feedback control section and a learning control section, which divides the engine's operating conditions into learning regions and updates air-fuel ratio learning values collectively across all regions if any region's learning is incomplete, allowing for earlier convergence to the target air-fuel ratio by reflecting the variation specific to the individual engine.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If learning is performed separately for each operating region, then the air-fuel ratio learning value can be accurately learned for each region, but the time required to complete learning of all regions is exponentially increased
Solution Approach 1:
The patent merges the learning processes of multiple operating regions by collectively updating air-fuel ratio learning values across all regions simultaneously. Instead of completing learning in one region before moving to the next, the system performs parallel learning updates across port injection and direct injection regions, significantly reducing the total learning time while maintaining accuracy through region-specific feedback control.
2Measurement precision
If the number of learning regions is increased to cover all operating conditions, then the air-fuel ratio control accuracy is improved, but the learning opportunities for each region are reduced and learning time is extended
Solution Approach 1:
The patent implements continuous learning updates across all operating regions simultaneously rather than sequentially. The learning control section continuously performs feedback control and updates learning values for both port injection and direct injection regions in parallel, ensuring that learning action is continuous and productive across all regions without idle waiting time.
Solution Approach 2:
The system prepares and maintains learning capability across all regions from the start, rather than activating learning in each region only when needed. By having the learning control section ready to update all regions simultaneously and using preliminary collective update logic, the system eliminates the sequential delay that would otherwise occur between regions.
3Reliability
If learning is performed from scratch in all regions after battery clearing, then complete and accurate learning values are obtained, but relatively long time is taken until efficient learning is achieved
Solution Approach 1:
After battery clearing, the patent merges the initialization process across all learning regions by performing collective updates from the start. Instead of initializing and learning each region separately from scratch, the system simultaneously begins learning in port injection and direct injection regions, combining their progress to achieve efficient learning much faster while ensuring complete coverage of all operating conditions.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
The air-fuel ratio feedback control section (120) updates an air-fuel ratio feedback correction value (FAF). The air-fuel ratio learning control section (130) performs, in each of learning regions (RP11-RP45), learning of an air-fuel ratio learning value. If the air-fuel ratio feedback correction value (FAF) converges to a value less than or equal to a specified value, the air-fuel ratio learning control section determines that learning of the air-fuel ratio learning value (KG) in the learning region has been completed. If it has not yet been determined that learning of the air-fuel ratio learning value (KG) has been completed in any of the learning regions (RP11-RP45), the air-fuel ratio learning control section collectively updates the air-fuel ratio learning values (KG) of all the learning regions at the time of updating the air-fuel ratio learning value (KG) through learning in any of the learning regions.