Abstract

BLANCAS-HERNANDEZ, Gerardo; LOBODA, Igor; GONZALES-CASTILLO, Iván  and  RENDON-CORTES, Karen Anaid. Comparison of fault severity estimation algorithms for gas turbine diagnostic systems. Ing. invest. y tecnol. [online]. 2022, vol.23, n.1, e1763.  Epub May 02, 2022. ISSN 2594-0732.  https://doi.org/10.22201/fi.25940732e.2022.23.1.001.

The objective of this study is to compare the estimation of failure severity using the two main approaches used for the diagnosis of flow duct in gas turbine engines. This study starts with the simulation of different faults of an engine by its nonlinear physics-based model (thermodynamoic model). The first approach uses simulated measurements and system identification techniques that estimate special fault parameters, which allow to localize faults and determine their severity. The second approach is based on pattern recognition theory and mainly uses data-driven models. The fault classification required for this approach can be made up of simulated patterns for each class of failure. In order to obtain safer and more general results, the comparison was carried out independently for three different classifications, each one having its own failure parameters. The results obtained show that in general the accuracy of the second approach is greater than that of the first, although for the last classification the accuracy of the approaches is comparable. The greatest difficulty that arose and also the greatest contribution of this work was determining how to estimate severity in the first approach. To solve this problem, we propose to use an artificial neural network. Another novelty is the comparison of two main approaches to diagnose gas turbines in the function of estimating the severity of failures. There are many studies whose main objective is to compare different techniques for diagnosing turbines, but none compare their severity estimation capabilities.

Keywords : Gas turbine diagnosis; failure severity; severity estimators.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )