A LPV adaptive observer approach to calibrate MAF sensor map in diesel engine

Abstract

In this paper, a novel method for compensating mass air flow (MAF) sensor error due to installation and aging is developed by using adaptive observer. Through introducing a concept of membership function, the relative error on the grid points is represented as piecewise linear interpolation model which is expressed as dot product between regression vector and parameter vector. With the combination of relative error model and air intake model of diesel engine, the MAF sensor relative error estimation is described as linear parameter varying (LPV) system with unknown parameter vector, and the corresponding LPV adaptive observer is designed to estimate states and parameter jointly. The exponential convergence of the algorithm is proved under the conditions of some persistent excitation and given inequalities. In order to correct MAF sensor error online, the estimated relative error is further described as a map. The observer performance is validated against the simulation data from engine software enDYNA. The results demonstrate that the proposed method can estimate relative error, and the true mass air flow can be estimated by MAF sensor measurements with the error map.