Métodos estatísticos aplicados à avaliação da qualidade da matéria-prima e classificação dos fornecedores de uma indústria de laticínios

Abstract

The aim of this study was to evaluate the quality of raw materials delivered by suppliers of a dairy industry based on physicochemical specifications. The development of the work consisted of literature research, followed by a case-study with descriptive approach involving variables of quantitative nature. In order to develop the analysis of data, the following statistical techniques were used: Exploratory analysis of data (used to evaluate the consistency of the collected data and to verify their behavior in relation to variability and normality), non-parametric methods (used to compare periods and collection patterns in relation to the behavior of the variables), CEP (used through the construction of control charts for attributes based on the specifications of the variables), multivariate analysis (used through the factorial analysis in the construction of indexes for the classification of suppliers). Two stages were observed in the development of the study. In the first stage, an analysis on the quality of milk based on the acidity determination was conducted. In the second stage, the analysis of the quality of raw materials was conducted using only lots with appropriate acidity values and considering the other physicochemical variables. The results obtained in the first stage were used to observe that most of the lots were produced according to quality standards established for acidity. For the comparison of the milk collection patterns, in general, pattern 1 obtained better performance in all seasons. In the second stage of the analysis, the results showed that suppliers presented lower percentage of milk lots with appropriate specifications for the percentage of water and density. Moreover, considering the comparison between the collection patterns in winter, it was possible to verify that patterns 2 and 3 had the best performance in the quality indicator. Considering the spring season, it was found that standard 1 presented the best performance. In addition, comparing the collection standards in the summer, it was found that standard 2 obtained the best performance. Through the control charts applied, it was possible to perceive that the process was out of control in the three periods mainly due to problems with the specification of the percentage of water and density. Finally, with the aid of the factorial analysis, indexes for the classification of suppliers were calculated based on physicochemical variables.