Identificación de un sistema de control de nivel y flujo utilizando redes neuronales artificiales

Abstract

In this investigation, it considered the identification of a level and flow control system through neural networks. The equipment is currently used for training activities in processes, instrumentation and control of variables. However, due to the absence of a process model, it is impossible to design a control system, the tuning of controllers is done directly. This also cause restrictions in tasks of detection, diagnosis and isolation of failures, area of current development in the industry and in the academy. Identification is the process of finding a process model without resorting to the laws of equilibrium but from data on the operation of the equipment. For this work, artificial intelligence was used to obtain the process model of level and flow control. The model obtained will be very useful for the design of controllers, detection, diagnosis, isolation and fault tolerance.