Balanceamento de linha de montagem de motores a combustão por meio de programação linear inteira mista

Abstract

In the search for productive excellence, the elimination of wastes as a way to increase the productivity is aimed. The problem of line balancing consists in the balanced distribution of tasks needed to manufacture a product along the workstations, reducing wastes by minimizing idle times of equipment and operators and the need of stocks. The goal of this project is to propose a line balancing solution to a segment of a combustion engines’ assembly line in an automotive company located in the metropolitan region of Curitiba, using a mathematical model in Mixed Integer Linear Programming (MILP), finding solutions that reduce idle time. To achieve so, a problem description was made, identifying details in the assembly line and its products. Afterward, a literature review involving the key definitions of line balancing was done. A methodology to achieve the goals of the project was then proposed. The development contains two steps involving: data collection; tasks grouping; restrictions identification; precedences diagram elaboration; feeding, solution, adaptation and validation of the mathematical model in MILP; results analysis; and, finally, presentation of a balancing proposal. In the context of results obtained, at a preliminary analysis, the methodology is applied on a single model line balancing problem, using the data of the model with greater demand. With the validation of the proposed methodology, the study proceeded with experiments involving a mixed-model line balancing, considering all models of engines produced. In the context of these experiments, an analysis of the terms needed for the mathematical model objective function was made, in order to acquire adequate results related to indicators of cycle time, idleness and smoothness indexes. The usage of a term in the global function, which gives relevance to the number of tasks relocated between the actual scenario and the proposed one, is also emphasized. Therefore, an optimized solution that aims less chances when compared to the present production line is pursued. The results of these studies have shown that there are significant potential gains in the line’s performance, as a reduction of 4.1% in cycle time, reduction of 43.5% in the total idle time and better tasks levelling between operators and workstations, taking the actions required for the implementation of the proposed solution in consideration. The results obtained enhance the possibility of a relative low cost and low complexity implementation that contributes importantly to the line balancing improvement.