Estratégia de identificação paramétrica aplicada à modelagem fenomenológica de um sistema do tipo correia transportadora industrial para fins de detecção de faltas

Abstract

Belt conveyor systems are essential for large companies, even though this equipment has a great deal of importance that makes its non-planned stop can generate huge amounts of losses or even the stall of the whole production process. Having this importance in mind it becomes necessary to realize the adequate monitoring of the system and detect with a larger prior notice the occurrence of some fault in the system. In an effort to reduce the unforeseen stop, this dissertation investigates a modeling of a belt conveyor system. At first a phenomenological model of the process is discussed based on the mechanical laws and considering the diverse types of movement opposition force throughout the belt conveyor. The main parameters of a belt conveyor belt where estimated through the non-recursive mean square. In sequence a fault detection algorithm was elaborated using the interval analysis theory, in a way that its possible to detect inadequate operation conditions. With the intent to evaluate the performance of the proposed algorithm a prototype that emulates the tipical operation of a conveyor belt was designed. The results were obtained experimentally, that confirms the great performance of the proposed methodology.