Aplicação da metodologia PDCA para redução da variabilidade na medição do parâmetro cor em resinas de PVC

Abstract

PVC, or Polyvinyl Chloride, is one of the most produced synthetic plastics in the world. Initially, it is obtained as a white powder without direct industrial use, requiring chemical additives to acquire properties such as flexibility or rigidity, essential for its various industrial applications. Each PVC resin has strict technical specifications, including parameters such as grain size, bulk density (BD), viscosity (VK), volatile materials (MV), thermal stability (ST), residual MVC (MVCr) and color. Color, crucial in quality assessment, is measured in terms of brightness ("L"), red/green ("a") and yellow/blue ("b"), with different classifications such as: crystal, transparent, opaque and off. The work was carried out in a real PVC factory, fictitiously called "PVCFlex", where the resin "SP 057RA", used in the manufacture of PVC liners, is analyzed daily by field operators and analysts from the Quality Control Laboratory (LCQ). However, significant discrepancies were observed between these measurements. The present work aimed to identify potential causes and reduce this variability. For this, the PDCA methodology (Plan-Do-Check-Act) was applied with the use of quality tools (Ishikawa diagram, cause and effect matrix, stress x impact matrix and 5W2H) and statistical tools (summary report, correlation analysis, regression equations and control charts). Thus, it was possible to identify the possible causes that contributed to the measurement difference, such as lack of standardization and violation of operational procedures, absence of calibration of equipment, differences between measuring equipment, among others. From the identification of these causes and prioritization of actions in the effort x impact matrix, an action plan was carried out that mainly contemplated the training of the team and the standardization of color analysis procedures and calibration of equipment. With these measures, a 20% reduction in the variability of color parameter measurement was achieved. Regression equations were also developed with the objective of converting the values obtained in the field to the equivalents measured in LCQ.