Methods and Times Measurement in Pasubio's Finished Selection Department

The Master Thesis originates from an internship experience at Pasubio S.p.A in Arzignano, a leading company in the production of leather for the high-end automotive sector. The focus is posed on the department, identified as the bottleneck of the production chain. The leather selection process proved to be a highly variable operation, with variability arising from the raw material, the production process, and the subjectivity of the operators’ evaluations. To address these challenges, a Time and Motion analysis represents the most appropriate approach, as it aims to standardize and optimize the production process. In addition, it allows for the definition of cycle times, thereby enabling continuous monitoring of efficiency and providing valuable support for production planning. The objective of this study is to evaluate the current state of the department with the aim of increasing awareness of its operating conditions and laying the foundations for a continuous improvement process. This analysis provides the essential basis for identifying potential inefficiencies and non-essential activities, thereby enabling targeted decisions to overcome them and ultimately improve the overall productivity of the department. The main activities carried out to achieve these objectives included the measurement and analysis of cycle times for the selection operation on a set of leather articles. Working times were recorded for multiple operators, across different shifts, days, and batches of leather, ensuring the representativeness of the data and allowing for the correction of the previously used target times. To validate the recorded data, a comparative test was conducted on a selection of leather batches (tags). This experiment confirmed the reliability of the measured times and revealed an 18% overestimation relative to the times recorded by the Manufacturing Execution System (MES), which the company currently uses to evaluate departmental efficiency. The analysis of the distribution of selection times revealed a positive skewness. A detailed investigation of ratio variables identified the lognormal distribution as the most representative probability model. On this basis, the estimation of the ideal sample size for each leather article was refined, correcting the overestimation of about 16% that results from assuming normality. Furthermore, the lognormal distribution was employed to fit a predictive model for estimating the mean selection times of tags, achieving a good fit as confirmed by the residual analysis. This approach integrates the general estimates at the item level with more detailed forecasts at the batch level. The definition of the predictive model, together with the exploratory analysis, made it possible to address the hypotheses formulated at the outset of the study and to identify the main drivers of selection time. Variables related to leather color and the distinction between Nappa and Stampa leathers, initially hypothesized as potential determinants, did not show a clear impact. In contrast, a substantial difference was observed comparing Nappa and Stampa leathers with Split leathers, approximately 1.7 in terms of selection time ratios. Split represent a fundamentally different product type, as they are leathers physically divided into two halves. The most influential factors were found to be the operators, reflecting the subjective nature of the operation, and the number and diversity of quality categories to be assigned, which highlight the influence of the product itself.