Knock Sensor Based Virtual Combustion Sensor Signal Bias Sensitivity

2018-01-1154

04/03/2018

Features
Event
WCX World Congress Experience
Authors Abstract
Content
The combustion in a direct injected internal combustion engine is normally open-loop controlled. The introduction of cylinder pressure sensors enables a virtual combustion sensor which in turn enables closed-loop combustion control, and the possibility to counteract effects such as engine part-to-part variation, component ageing and fuel quality diversity.
Closed-loop combustion control requires precise, robust and preferably cheap sensors. This paper presents an investigation of the robustness and the limitation of a knock sensor based virtual combustion sensor. This virtual combustion sensor utilize the common heat release analysis using a knock sensor based virtual cylinder pressure signal.
Major virtual sensor error sources in a heavy-duty engine were identified as: the specific heat ratio model, the boost pressure and the crank angle phasing. The virtual sensor errors were quantified in relation to both the measured cylinder pressure and the total virtual sensor error.
The tolerance analysis of the virtual sensor showed the signals of the crank angles 10% and 50% heat release as robust with low sensitivity to errors. An additional dependency found for the crank angle of 10% heat release was the compression ratio error.
The study concluded that it is possible to estimate the mass fraction of 10% and 50% heat release accurately enough to make the virtual sensor an option for closed loop combustion control.
The study was not able to prove the virtual sensor cylinder pressure signal, crank angle of burned fractions signals or total heat release signal capable of confident determination of the amount of biodiesel in fossil diesel.
Meta TagsDetails
DOI
https://doi.org/10.4271/2018-01-1154
Pages
13
Citation
Rugland, C., and Stenlaas, O., "Knock Sensor Based Virtual Combustion Sensor Signal Bias Sensitivity," SAE Technical Paper 2018-01-1154, 2018, https://doi.org/10.4271/2018-01-1154.
Additional Details
Publisher
Published
Apr 3, 2018
Product Code
2018-01-1154
Content Type
Technical Paper
Language
English