Análise de variância fatorial aplicada na avaliação da resistência de argamassas industrializadas

Abstract

The analysis of experiments has been used by industries that search for the continuous improvement of processes and products, in other words, those effectively involved in assuring the satisfaction of their customers. Industrial experiments frequently require studying the effect of one or more factors; therefore, it is verified that the number of tests trends to increase as the number of factor grows, making this task exhaustive and quite expensive for companies. On the other hand, the combinations of the experiment factors need to be analyzed in order to verify its validity. One of the solutions for this situation is to apply the factorial experiments technique. The objective of the present work was to perform a study on the factorial analysis of variance and comparison of averages in order to apply and statistically compare the compression strength of the different industrialized mortar traces produced at different shifts at a given company. To do so, this methodology was applied in the mortar industry, and the data were supplied from the Santa Maria Federal University tests laboratory and evaluated through the performance of factorial analysis of variance at significance level of 5% as soon as the results that compose the analysis of variance were determined, so averages of strengths that the mortar traces present could be compared two-by-two. Thus, it was identified that there was no interaction between trace and shift and that the difference found between traces was significant, unlike the difference between shifts that was not, and the descriptive analysis of data indicated that the average strength is above recommended values, which could be treated for an adjustment in the productive process to occur in order to show the closest averages in relation to the specified values, reducing the variability around them, generating products with fine quality and adjusted production costs. No significant differences between the production shifts were observed.