*5.3. Data Collection*

Times were taken, specifically 73 records in each operation during February to April 2019, considering the hours when there was low and high production. Neither the first nor the last hour of the shift were used. The data was checked for detecting and discarding atypical data. These data were entered in the input analyzer, which is a tool of the Arena™ software. This allows determining which specific distribution fits the data, and which uses the chi square χ<sup>2</sup> and the Kolmogorov-Smirnov test as goodness-of-fit. The distributions that accept the null hypothesis in were used for obtaining the different distributions of each activity, which will be used to simulate the process. Figure 8 shows the analysis of the flaming operation, including the times taken, the analysis of fit test, and the expression that was used to model the production process in Arena™.

**Figure 8.** Probability distribution found for the flaming activity, elaboration in the Arena™ software.

Proposal: For doing the simulation model, the same distributions of the operations that were calculated with the times collected were considered.
