Engine Torque Estimation Using Nonlinear Kalman Filter Noise Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge is to accurately estimate the indicated torque of each cylinder in an engine due to limitations in crank angle sensor measurement resolution, sampling period, and engine vibrations, which hinder effective feedback control.
Innovation Solution
An engine torque estimating apparatus and method using a microcomputer that processes crank angle and angular velocity data, employing a nonlinear Kalman filter to calculate estimated torques, and acquiring conversion maps to suppress noise and improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If feedback control is implemented using crank angle sensor values to estimate indicated torque of each cylinder, then individual cylinder torque control is enabled, but measurement noise from the crank angle sensor prevents accurate torque estimation
Solution Approach 1:
The patent introduces an intermediary processing system (noise suppression processing) between the crank angle sensor and the torque estimation algorithm. This intermediary component filters measurement noise while preserving the underlying torque signal, enabling accurate torque estimation despite sensor limitations. The noise suppression processing acts as a mediator that reconciles the conflicting requirements of using sensor data for control while eliminating its harmful effects.
Solution Approach 2:
The patent implements feedback control where the estimated indicated torque of each cylinder is used to adjust injection amounts. The noise suppression processing enhances this feedback loop by providing cleaner torque estimates, allowing the system to effectively counteract individual cylinder torque variations and achieve precise individual cylinder control despite the noisy sensor inputs.
2Speed
If crank angle sensor data is used for torque estimation, then real-time control is enabled, but sensor limitations (measurement resolution, sampling period, vibration) introduce noise that degrades estimation accuracy
Solution Approach 1:
The patent converts the harmful measurement noise into a manageable signal by applying noise suppression processing that distinguishes between high-frequency noise components and the lower-frequency torque signal. The system accepts the noisy sensor data as inevitable but transforms its harmful effect into a beneficial filtered signal that retains the essential torque information while eliminating vibration and resolution-related noise.
Solution Approach 2:
The patent changes the parameters of the signal processing by applying noise suppression filters that selectively attenuate certain frequency ranges while preserving others. This parameter-based approach allows real-time processing by transforming the raw sensor signal into a cleaned version that maintains temporal resolution while improving accuracy through selective frequency modification.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An engine torque estimating apparatus having a processor and a memory accessed by the processor. The processor performs a torque estimating that calculates time series data of an estimated indicated torque, based on a crank angle having been extracted from an output of a crank angle sensor of an engine including a plurality of cylinders; an estimated indicated torque-related value extracting that extracts an estimated indicated torque-related value, for each of the cylinders, from the time series data of the estimated indicated torque, for each of the cylinders; and average indicated torque correct value acquiring that converts, for each of the cylinders, the estimated indicated torque-related value into an average indicated torque correct value having been calculated based on a cylinder internal state of the engine in correspondence to the estimated indicated torque-related value.