Tratamento de dados em sistemas de automação de nível superior para cálculo de indicadores

Abstract

With the use of automated systems and the increasing volume of data generated in an organization, the error in collecting, processing the data or the calculation of an indicator generate incorrect values, which may encourage inappropriate management decisions, steering the organization of their goals. To correct these errors, or reduce its impact on the calculated value, a sort of methodologies can be used, each methodology applied in a specific case. This paper presents a review of data processing techniques for the calculation of indicators in automation systems level. Among the methods of treatment of errors in the calculation of the indicators, is mentioned: down sampling, filtering, interpolation, study of error propagation and statistical methods. It is considered that evaluation of errors related to the calculation of the indicator should be investigated during the creating of the indicator, seeking to identify and minimize calculation errors. The process of definition and calculation of indicators is also examined. The techniques are illustrated with application data from real processes.