Apoio à tomada de decisão e minimização da perda de matéria prima em processos de manufatura

Abstract

The following work aims to optimize and make automated a manufacturing system that uses metal plates as raw material. The analysed process currently has several steps with no communication between each other, moreover requires a high level of user’s intervention in their decision making. This process must be unified in order to achieve the optimal result based on the optimization criteria proposed. The analysed processes throughout this work address all stages of the manufacturing process involving plates, going through the initial stages as the definition of the parts to be produced, intermediate steps as the process of optimizing parts (nesting) and its parameters and final step, i.e., obtaining the CN code for the production of parts. To optimize these steps, a method of hybrid solution using a system to support the decision making aided by an expert system and known optimization (such metaheuristics) is proposed. The proposed method will improve the results through an automated parameterization using the system to support the decisions making, defining the best parameter based on the products to be manufactured, thus reducing the need for manual decisions and therefore the user input in the process and eliminating the need for an expert to be the same. This automation should analyse the possible plates in stock as well as parameters offered by the nesting process and seek the best configuration for the process, analyzing possible permutations. Furthermore, we propose the addition of a new component in the steps of manufacturing, responsible for analysing the process remains usable and organize the scrap generated by it, making it available for reuse in the future. Through the proposed system, good and relevant results were obtained. In several cases the obtained results are better than results in the literature. In addition, the acceptance of the system by users who performed performance system tests was considered excellent. As pointed by these users, the number of parameters to be selected has drastically reduced, thus making the system simpler to use.