Sistema de inferência fuzzy para classificação de impactos ambientais gerados por postos revendedores de combustíveis

Abstract

A fuzzy inference system (FIS) to assess the environmental impact of contamination level generated by leaks in Stations Fuel Dealers (SFD) was developed in this work. An Environmental Impact Index for Leak in Stations Fuel Dealers (EIIL-SFD) were obtained using seventeen input linguistic variables: type of contaminant, leaked volume, impacted area, the contaminant physical condition, existence of containment barriers, topography land, greenhouse gas emissions, proximity to rivers/streams, proximity of wells, thickness of the aquifer, annual precipitation, flood potential, proximity to residential areas, proximity to shopping areas, proximity to rural areas and agribusiness, proximity to schools and/or nurseries and proximity to areas of environmental preservation - grouped in partial indexes, according to the source of the contaminant, the propagation of the contaminant and the site of contamination. The fuzzy inference Mamdani method was used for mapping input and output linguistic variables, using a base composed of 112 rules, based on expert knowledge and triangular membership functions. The higher the EIIL-SFD which range between 19.9 and 100 - greater the degree of contamination of the study area, indicating a greater urgency in decision-making on intervention/remediation of the contaminated area. The SIF developed was used successfully in case study of 3000-liter diesel fuel leak, due to the pipeline disruption of the underground fuel supply system of a SFD, located in the metropolitan region of Recife, state of Pernambuco, Brazil, presenting EIIL-SFD equal 63.7 - classified as serious environmental damage with a fine of application. The results suggest that the IIAV-PRC can be used - by regulatory agencies and/or consultants and / or tenants and / or owners and / or distributors - as a support tool for the assessment of environmental impact generated by leaks in PRC and the taking decisions about remediation actions of impacted areas.