Abstract
SUMMARYIn this paper, we propose a fault detection and isolation filter design method for internal combustion spark ignition engines. Starting from a detailed nonlinear mean‐value mathematical description of the engine, a novel linear parameter varying (LPV) model approximation is derived on the basis of a judicious convex interpolation of a family of linearized models. A filter structure consisting of a bank of LPV observers is considered, each of them in charge of detecting a particular class of faults and exhibiting low sensitivity to all other faults and exogenous inputs. The resulting diagnostic filter is parameter‐dependent in that a set of measurable engine variables is used online to suitably modify the filter gain so as to better take care of system nonlinearities. The quality of the LPV model approximation of the engine and the diagnostic capabilities of the fault detection and isolation architecture are demonstrated by a series of extensive numerical simulations. Copyright © 2013 John Wiley & Sons, Ltd.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call