Engine Torque Estimation Using Error Learning and Angular Acceleration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing torque estimation methods for internal combustion engines struggle to accurately account for individual variations and aging changes due to limitations in the learning capabilities of torque converters.
Innovation Solution
A torque estimation apparatus that includes a rotation information detection unit, first and second estimation torque calculation units, and a third estimation torque calculation unit to learn and correct torque variations using angular acceleration and error learning values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If torque converter characteristic is used for learning output torque variation, then individual difference and aging change of internal combustion engine can be learned, but learning accuracy is limited due to variation in torque converter characteristic and limited operating region
Solution Approach 1:
The patent divides torque estimation into multiple independent calculation paths: a first estimation torque calculation unit using engine characteristics (intake air amount, fuel injection amount, ignition timing), a second estimation torque calculation unit using torque converter characteristics, and a third estimation torque calculation unit that combines both. This segmentation allows each path to operate independently with its own strengths, improving overall estimation accuracy without being limited by the weaknesses of any single method.
Solution Approach 2:
The patent merges the first estimation torque (based on engine characteristics) and the second estimation torque (based on torque converter characteristics) through a synthesis unit that calculates a synthesis coefficient. This combination leverages the complementary strengths of both methods: engine characteristics provide accurate baseline torque estimation, while torque converter characteristics capture variation due to individual differences and aging, achieving higher overall accuracy than either method alone.
2Measurement precision
If torque converter characteristic is used for learning, then torque variation can be learned, but the operating region is limited
Solution Approach 1:
The patent creates a universal torque estimation system that can operate across all engine conditions by making the first estimation torque calculation unit independent of torque converter operating region limitations. This unit uses fundamental engine parameters (intake air amount, fuel injection amount, ignition timing) that are valid across the entire operating range, while the second estimation torque calculation unit handles specific regions where torque converter data is available. The synthesis unit combines both, ensuring the system works universally across all operating conditions.
3Measurement precision
If first estimation torque is calculated using torque characteristic data, then baseline torque can be estimated, but individual variation and aging change are not reflected
Solution Approach 1:
The patent implements a feedback mechanism where the second estimation torque calculation unit continuously monitors actual torque converter behavior and calculates a synthesis coefficient that reflects individual differences and aging changes. This synthesis coefficient is fed back to the third estimation torque calculation unit, which uses it to correct the first estimation torque. This closed-loop feedback ensures that the baseline torque estimation is continuously adjusted to reflect actual engine condition changes, maintaining reliability over time.
Data Source
AI summary
To provide a torque estimation apparatus for an internal combustion engine which can learn the variation in the output torque of the internal combustion engine caused by the individual difference and the aging change of the internal combustion engine with good accuracy. A torque estimation apparatus for an internal combustion engine calculates a first estimation torque using a torque characteristic data for first estimation torque; calculates a second estimation torque based on an angular acceleration; learns a deviation between the second estimation torque and a preliminarily set reference value of the second estimation torque, as error learning value; and calculates a value obtained by correcting the first estimation torque by the error learning value, as the third estimation torque.