Data integration: Quality aspects

It is now possible to observe that the advances in technology have led to increased capacity to generate and store huge amounts of data in all areas of knowledge, characterizing a generalized explosion of data. In the area of electric power, the vast amount of data collected by different systems of supervision and control and stored in historical bases, has become a potential source of implicit knowledge of great value to the management of electrical systems. The availability of huge volumes of data, from different sources with different levels of quality, makes the extraction of this knowledge time consuming, costly and often ineffective. Additionally, there is the need to evaluate the quality of primary data between those historians, since they may present incomplete or inconsistent due factors related with industrial processes of acquisition: Measuring instruments with bias, uncalibrated or malfunctioning: Manual entries of incomplete data, illegible or not made: Stops and starts of equipments: Abnormal behavior of the process In this paper we consider the search for quality in the user's perspective of the information ("fitness for use" principle) where the definition of the level of an adequate quality of data depends on the context. We seek to verify if information currently existing in systems providing data are sufficient to ensure the quality of their data. This check is made by mapping the information in metrics of the quality of data. For this study, an adaptation of a methodology was done with the following steps: identification of the system being evaluated; identification and selection of the most important data sets; evaluation of the quality of selected system and finally, the quality measurement of their data. The first step was characterized by the choice among the various systems, the system being the object of evaluation, considered in this study as the Events Repository System. This system consists of a set of computational modules to support the tasks of analysis and monitoring of large industrial processes. The quality step of the information system focused verifying the functionality of this system in the aspects of data storage and retrieval. Several dimensions of quality relevant to the intended application of data, were defined and assessed in this step. With these dimensions defined it was possible to check the quality of the stored data, and point out the actions and precautions to be taken to ensure the quality that is required.