Fuel Injection Controller Learning Speed Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing fuel injection controllers face challenges in setting appropriate fuel injection amounts due to influences from short-term, medium-term, and long-term fluctuations in engine states, leading to inefficient fuel control and potential errors in air-fuel ratio management.

Innovation Solution

A fuel injection controller that utilizes an oxygen sensor to determine an injection amount correction value, with separate short-time and long-time learning values updated at different speeds to isolate and manage these fluctuations, ensuring accurate fuel injection control even during interruptions in feedback control.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If the learning speed is increased to make the learning value follow the feedback correction factor quickly, then the response speed improves, but the learning value becomes more easily influenced by short-term and medium-term fluctuations of the engine state

Engineering Contradiction:
Improvelearning speedVSAvoidaccuracy of fuel injection control
Core Design Contradiction:
SpeedVSReliability

Solution Approach 1:

The patent divides the single learning value into two separate learning values: a first learning value updated at a first learning speed and a second learning value updated at a second learning speed. This segmentation allows the system to handle different time scales of engine state fluctuations separately, with the first learning value responding to short-term changes and the second learning value capturing long-term trends, thereby resolving the contradiction between response speed and control accuracy.

Inventive Principle:
Principle #1Segmentation

2Stability of the object's composition

If the learning value is set to absorb only long-term fluctuations, then the stability improves, but the feedback correction factor must accommodate both short-term and medium-term fluctuations, causing insufficient follow-up performance when feedback control is interrupted

Engineering Contradiction:
Improvestability of learning valueVSAvoidfollow-up performance
Core Design Contradiction:
Stability of the object's compositionVSProductivity

Solution Approach 1:

The patent segments the learning function into two distinct components with different update speeds. The first learning value (updated faster) absorbs short-term and medium-term fluctuations, while the second learning value (updated slower) captures long-term trends. This segmentation enables the system to maintain both stability and responsive follow-up performance, as each learning value handles specific time-scale fluctuations appropriately.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements dynamic learning speeds by updating the first learning value at a first learning speed and the second learning value at a second learning speed. This dynamic approach allows the system to adaptively respond to different operational conditions, maintaining high follow-up performance during transient states while ensuring stability during steady-state operation.

Inventive Principle:
Principle #15Dynamics

3Device complexity

If a single learning value is used, then the device complexity is reduced, but the system cannot distinguish between short-term, medium-term, and long-term fluctuations, leading to inappropriate fuel injection control

Engineering Contradiction:
Improvestructure of learning systemVSAvoiddetection of engine state fluctuations
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent applies segmentation by dividing the learning system into two distinct learning values with different update characteristics. This segmentation enables the system to detect and respond to different time scales of engine state fluctuations (short-term, medium-term, and long-term) separately, improving measurement precision without requiring complex additional hardware, only computational differentiation.

Inventive Principle:
Principle #1Segmentation

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach improves fuel efficiency and exhaust cleanliness by accurately managing fuel injection amounts, reducing the impact of short-term and medium-term fluctuations and ensuring stable control before and after feedback control is resumed.

Implementation Method 1

an oxygen sensor that responds to an oxygen concentration inside an exhaust passage through which an exhaust of the engine passes

Methodology Applied
Scientific EffectOxygen sensing:

Data Source

PatentEP2796693B1Fuel injection controller
Publication Date: 2017.06.21 YAMAHA MOTOR CO LTD
  • EP2796693B1 patent drawingFigure 1
  • EP2796693B1 patent drawingFigure 2
  • EP2796693B1 patent drawingFigure 3

AI summary

A fuel injection controller includes an oxygen sensor (33) for responding to an oxygen concentration inside an exhaust passage (43), and an injection amount control unit (50) programmed to control a fuel injection amount based on the output of the oxygen sensor (33). The injection amount control unit (50) includes an injection amount correction value computing unit (66) for determining an injection amount correction value (C) based on the output of the oxygen sensor (33), a short-time learning value computing unit (67) for determining a short-time learning value (S) based on the injection amount correction value (C), a long-time learning value computing unit (68) for determining a long-time learning value (L) based on the short-time learning value (S); a feedback correction amount computing unit (65, 71) for computing a feedback correction amount, an injection amount control value computing unit (69) for computing a control value of the fuel injection amount, and a long-time learning value holding unit (52N) for holding the long-time learning value (L).