Numerical solution for finding time constants in Harriot's method for identifying overdampened second-order systems

Abstract

The identification of systems seeks to use techniques that, from input and output signals, can find a dynamic model that describes the system. Among the simplest identification techniques are those based on step response. In this work, the Harriot method is approached, a graphical method that identifies over dampened second-order models. This study presents a numerical solution to find the time constants of the model, allowing its application in a computational and automated way. To validate the proposed approach, it was applied in a motor-cogenerator didactic module and compared with other deterministic identification techniques, such as Ziegler-Nichols, Hägglund, and Sundaresan, to compare the results. The results showed that the numerically resolved Harriot method performed equivalently or even better compared to the other deterministic identification techniques. The R² and Akaike Information Criterion (AIC) were used as metrics to quantify the performance of each model. It was concluded that the numerical solution proposed for the Harriot method allows its computational application and presents a promising performance. Future work may explore the use of this technique as part of an auto-tuning scheme for PID controllers.

DOI:https://doi.org/10.56238/alookdevelopv1-117