Modelagem e classificação de cortes de serras feixe circulares industriais com foco em confiabilidade

Abstract

In industrial environments, the correct definition of operating parameters is fundamental to guaranteeing production rates and asset conservation. The equipment in this study consists of 4 circular saws designed to cut circular tubes. The cutting process on circular saws is influenced by the dimensions of the material and its chemical composition, as well as the characteristics of the blade, such as the number of teeth and diameter. In addition to the controllable variables, the sawing process is also affected by the conservation of the asset and the efficiency of the lubrication and cooling processes. In this way, defining the best cutting speed can become a complex task, given the large number of variables capable of influencing it. The equipment in question had outdated operating standards. Thus, in new product scenarios, operators set cutting speed and feed values based on their own experience. This situation led to forced deterioration, increasing the occurrence of breakages and wear on the cutting blades. The aim of this work was to use historical data from the PIMS - Plant Information Management System - to evaluate, using supervised learning, which combinations of cutting speed and feed are the most suitable for new products, thus revising the current procedure. To classify the data, c-means algorithms were used to group together steels of similar hardness and fuzzy c-means to evaluate cutting performance. After revising the operating standard, the MTBF (Mean Time to Failure) and MTTR (Mean Time to Repair) indicators improved substantially, which can be translated into a reduction in operating costs.