Air Intake Valve Model Learning for Engine Air Quantity Calculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for calculating air quantity in internal combustion engines face inaccuracies due to manufacturing tolerances and durability decline, leading to reduced learning accuracy of the air intake valve model, as the air intake amount detected by flow meters does not accurately represent the actual air flowing into the combustion chamber.
Innovation Solution
An air quantity calculation device that uses an air intake valve model to simulate air flow into the combustion chamber, incorporating an air flowing amount calculation unit, a determination unit to verify the match between calculated and actual air intake amounts, and a learning unit to correct model characteristics based on comparisons, ensuring accurate learning and correction of air quantity calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If air intake amount is calculated based on air flow meter detection value, then air quantity can be measured, but learning accuracy degrades when air flows downstream of the detection location
Solution Approach 1:
The patent introduces a determination unit as an intermediary that verifies whether the air intake amount from the flow meter matches the actual air amount calculated by the air intake valve model. This intermediary layer filters out unreliable data where air flows downstream of the detection location, ensuring only valid data pairs are used for learning, thus resolving the contradiction between measurement capability and learning reliability
Solution Approach 2:
The patent implements a feedback mechanism where the determination unit continuously monitors and compares the air intake amount from the flow meter with the actual air amount from the model. When a mismatch is detected (indicating air flowed downstream), the system feedbacks this information to exclude the data from learning, creating a closed-loop system that maintains learning accuracy while preserving measurement functionality
2Measurement precision
If air intake valve model is learnt using detected air intake amount, then calculation accuracy can be improved, but manufacturing tolerances and durability decline cause characteristic deviation
Solution Approach 1:
The patent applies parameter changes by introducing a learnt value that adjusts the characteristic values of the air intake valve model. Instead of relying on fixed manufacturing specifications, the system dynamically modifies model parameters (such as valve lift characteristics) based on actual operational data, thereby compensating for manufacturing tolerances and durability decline to maintain calculation accuracy
Solution Approach 2:
The system performs self-correction through the learning unit, which automatically adjusts the air intake valve model characteristics based on comparisons between detected and calculated air amounts. This self-service mechanism enables the system to compensate for its own manufacturing deviations and degradation over time without external intervention, maintaining accuracy despite tolerance variations
Data Source
AI summary
An ECU calculates a cylinder flowing air amount based on an intake tube pressure, using an air intake valve model simulating a behavior of air flowing into a combustion chamber via an air intake valve. The ECU calculates an air intake amount based on a detection result of an air amount detection sensor, and determines whether the air intake amount matches an actual amount of air flowing into the combustion chamber. When the air intake amount is determined to match the actual amount, the ECU calculates a learnt value based on comparison of the cylinder flowing air amount with the air intake amount.