Decentralized identification for multivariable integrating processes with time delays from closed-loop step tests

In order to identify those multivariable processes with integrating factors in their transfer function matrices, a simple yet robust decentralized identification method from closed-loop step tests is proposed. By the frequency response matrix computed from the closed-loop system data and the knowledge of the decentralized controller, the structural information of the multivariable integrating process is determined firstly and then the continuous parametric model with dead times is approximated similarly with the parameterization of the open-loop stable process. Computer simulations and an application to a 3 x 3 integrating multiple-tank water level system verify the validation of the proposed method even if the closed-loop system is affected by some stochastic noise sources.