Heat transfer coefficient estimation of an internal combustion engine using particle filters

The continuous growth of the worldwide vehicle fleet is leading to many private initiatives and governmental policies for the achievement of lower pollutant and greenhouse gas emissions through energetic efficiency improvement. The in-cylinder heat transfer has been generally recognised as a major factor influencing internal combustion engines efficiency and exhaust emissions. Therefore, the analysis of in-cylinder heat transfer is of importance for fuel economy, as well as for environmental preservation. In the present paper, the use of particle filters applied on the time-dependent heat transfer coefficient estimation of a combustion process in a spark-ignition internal combustion engine is presented. The combustion process was modelled as a transient zero-dimensional process based on the First Law of Thermodynamics for a perfect gas, and the fuel burned fraction was modelled through a Wiebe's function model. Results with different measurement frequencies and errors are available, using a sample importance resampling filter, showing a very good estimate, even for initial guesses far away from the real value of the unknown function.