Unbalance identification for a practical turbofan engine using augmented Kalman filter improved with the convergence criterion

Kalman filter has emerged as a powerful tool for unbalance identification in rotating machinery. Recently, the augmented Kalman filter combined with the finite element model has grown up and projects its potential for complex rotor systems. This paper investigates the application of the augmented Kalman filter (AKF) to a practical turbofan engine. The current study reveals that using steady-state responses as measurements can cause fluctuation in the estimated results, even divergence for some cases, while the available signals in practice are steady-state responses generally. To the authors' knowledge, this practical problem is revealed for the first time. To address the problem, the convergence criterion is employed to improve the AKF and formulates the adaptive fading augmented Kalman filter (AFAKF) proposed in this paper. Results indicate that the increase of the amplification factor, the insufficient measurement points, and the complexity of the dynamic model can all lead to the deterioration of the estimated unbalance. The proposed AFAKF method shows favorable convergence and can achieve accurate estimation with less than 5% relative errors, and the superiority over AKF in computation cost is also observed.