Real-Time Acceleration Harmonics Estimation for an Electro-Hydraulic Servo Shaking Table Using Kalman Filter With a Linear Model

Abstract

For an electro-hydraulic servo shaking table, there are usually various nonlinearities, which cause harmonic distortion of the sinusoidal shaking tests, lowering the control performance. Based on a Kalman filter, an acceleration harmonic identification scheme is developed to provide harmonic information. A linear state space model of the sinusoidal acceleration response is first built. The Kalman filter is employed to estimate the state of the linear model. The harmonics, including the fundamental response, can be directly estimated online, and the amplitude and phase of each harmonic can also be computed. As it eliminates the need for restoring the harmonics, the developed harmonic estimation algorithm is computationally more efficient than computing the harmonics indirectly from the estimated amplitudes and phases. What's more, the state transition matrix and measurement matrix are all constant matrices. The efficiency, accuracy, and real-time performance are demonstrated by experiments. Last, it is compared with another identification scheme that we developed to further stress its novelties.