StepID - A Matlab-based toolbox for identification fromstep response

Abstract

Abstract For advanced control, transfer function models are widely used and identification from step response is a common choice in process industries to obtain such models. Over the last decade, significant developments have taken place in the field of identification from step response using the so-called integral equation approach. However, these developments have not been included in the identification tools available commercially. This work focuses on development of a MATLAB toolbox in the form of a graphical user interface (GUI) that will be a specialized tool for identification from step response. Due to the recent advancement in data storage capacity, industries have become repositories of enormous amounts of data. However, proper tools are needed to extract valuable information from raw data. The development of the toolbox started as an Undergraduate Research Experience Program (UREP) project from the 5th cycle and it was further modified and extended in the 7th cycle. The functionality of the toolbox includes data import from an excel file or a MAT file, preprocessing of data and the use of different methods to estimate continuous time transfer function model parameters of different orders under different conditions of the process. The toolbox includes methodologies to estimate model parameters and time delay simultaneously under steady or unsteady initial conditions using least squares and instrumental variable method. Methodologies to handle higher order models and non-minimum phase processes are also included. Finally, the toolbox has the capability to choose the best model from a list of models obtained using different methods.