Fractional-order model identification based on the process reaction curve: A unified framework for chemical processes

Abstract

This study introduces a novel method for identifying dynamic systems aimed at deriving reduced-fractional-order models. Applicable to processes exhibiting an S-shaped step response, the method effectively characterizes fractional behavior within the range of fractional orders (α∈[0.5,1.0]). The uniqueness of this approach lies in its hybrid nature, combining one-variable optimization techniques for estimating the model fractional order α with analytical expressions to estimate parameters T and L. This hybrid approach leverages information from the reaction curve obtained through an open-loop step-test experiment. The proposed method demonstrates its efficacy and simplicity through several illustrative examples, showcasing its advantages over established analytical and optimization-based techniques. Notably, the hybrid approach proves particularly advantageous compared to methods relying on the process reaction curve. To highlight its practical applicability, the identification algorithm based on this hybrid approach is implemented on hardware using a microprocessor. The experimental prototype successfully identifies the First-Order Plus Dead Time (FFOPDT) model of a thermal-based process, validating the proposed method's real-world utility.