Improving residual generation for OBD in the air path of a truck engine Berton, Antoine
Publication details: Linköping Linköping University. Department of Electrical Engineering. LiTH-ISY-R-2648, 2003Description: 35 sSubject(s): Online resources: Abstract: A method to improve model based diagnosis for the air path of a truck engine is presented. Originally, inaccuracies in both volumetric efficiency model and sensors limited the diagnosis performances. Two approaches are presented to improve the diagnosis performances. The first method is based on an elimination of part of the data according to the level of a certain operational variable. The second and preferred method make use of adaptive statistical charts built to generate residuals which show a reduced dispersion. Data taken from two different trucks working in various operating conditions were used to evaluate the proposed approaches. It was shown that the use of charts improved the diagnosis performances by approximately 25 %.A method to improve model based diagnosis for the air path of a truck engine is presented. Originally, inaccuracies in both volumetric efficiency model and sensors limited the diagnosis performances. Two approaches are presented to improve the diagnosis performances. The first method is based on an elimination of part of the data according to the level of a certain operational variable. The second and preferred method make use of adaptive statistical charts built to generate residuals which show a reduced dispersion. Data taken from two different trucks working in various operating conditions were used to evaluate the proposed approaches. It was shown that the use of charts improved the diagnosis performances by approximately 25 %.