Adaptive integral filter for the identification of two-input two-output processes

Abstract

Measurement noise and loop interaction introduce errors during the identification of a two-input two-output (TITO) process. This paper proposes a novel method for the identification of a TITO process in the presence of measurement noise and loop interaction. The proposed structure includes two preload relays (P_relays) in the forward paths and two adaptive integral filters in the feedback paths to induce the limit cycle outputs. The integral filters are designed adaptively during the identification. The integral filters in the feedback paths reduce the interaction between the loops and induce denoised limit cycle outputs. Then a pair of parametric transfer function models for the TITO process dynamics is obtained using the limit cycle information. Also, a new approach to estimate the steady state gains of the models is proposed. Simulation examples illustrate the effectiveness and the simplicity of the proposed identification and control method.