Indicators of air quality as a tool for environmental management in operation of thermal generating units with natural gas: a case study of the Mario Lago Power Station

Abstract

This work intended to implement a research based on current acquisition practices for air quality data from a typical thermoelectric natural gas power plant – Usina Termelétrica Mario Lago – searching information about the relationship levels concerning energy generation, air quality and meteorological measurements. Thus, the study provides indicators related to the company operation, consequent exhaust gas emission, air quality, so that meteorological interference, and the company generation itself, could have their contribution considered. A bibliographic review, especially those related to environmental indicators and Data Mining techniques is presented. Two main databases were analyzed: one containing hour to hour data from twenty aeroderivative turbines covering 2002 through 2005, and another responsible for air quality – for two shelters located close to the power plant, and meteorological data – for three locations close to the company for the same period of time. Data mining was performed using decision trees generated by the software WEKA (Waikato Environment for Knowledge Analysis), a set of algorithm implementations using several data mining techniques. It is a descriptive research as it aimed at analyzing real data using data mining and supplying result interpretation focusing on causes and consequences. Furthermore, it is an applied research as well, since the final objective was proposing indicators that allow direct visualization of the operational influence of the company on air quality.