Detección y diagnóstico de fallos en sensores de temperatura en una columna de destilación fraccionada

Abstract

This article presents the design of a fault diagnosis and isolation system on sensors for temperature measuring on a fractional distillation column to prevent unstable operation. The diagnosis and fault isolation system is based on a state estimation model using state observers. The distillation column is modeled by a nonlinear mathematical model deduced from the material and component global balance for a binary mixture (alcohol-water), the model describes the behavior of the liquid molar concentration of the more volatile component of the mixture, also it allows to establish an operating point on the system which is used as a reference for a secure, stable and linearly predictable operation. The system includes the design of a set of reduced order observers which besides operating as virtual sensors are also used to generate residues, these are used to diagnose faults in the temperature sensors. The proposed fault diagnosis and isolation system for the distillation column performs residues generation and evaluation for fault detection, diagnosis and isolation to be performed in a short period of time.