Stochastic model for setpoint of a rolling mill: an application in the soybean oil production

Abstract

The present paper proposes a stochastic model for soybean oil extraction, introducing an automated and reliable method for flake thickness control. In this production process, to have a robust solution, we must consider the uncertainty arising, e.g., from inaccurate sensor readings and unforeseen changes in production. The main objective is to maximize the oil extraction by keeping the thicknesses of the soybean flakes within an operating range. Data collection performed in the industrial plant made it possible to analyze the input data with their respective outputs. A methodology based on multiple linear regression was used to describe the relationship between the variables. We propose a mathematical stochastic model to obtain pressure setpoints that produce flakes as ideal as possible for oil extraction. The results reported time and quality improvements within the production process.