Método para diagnóstico de etanol adulterado em motor de combustão interna em funcionamento através da vibração

Abstract

All organic material generates energy, but biofuel is manufactured on a commercial scale from agricultural products such as sugarcane, in the case of ethanol alcohol. Currently, Brazilian ethanol represents the best and most advanced option for the sustainable production of biofuels on a large scale in the world. The use of fuel ethanol reduces the emission of greenhouse gases and reduces energy dependence on fossil fuels. However, even with inspection by the National Agency of Petroleum, Natural Gas and Biofuels (ANP) at gas stations, cases of tampering are not rare. In the case of ethanol, common adulterations include the mixture of hydrated ethanol with anhydrous ethanol (the one that is mixed with gasoline, which is orange in color), methanol, or even water. Such tampering is visually imperceptible, but can cause changes in the car's performance, in the concentrations of pollutant gases produced and in the levels of deterioration of parts, generally leading to high costs for the repair of components affected by the damage caused by tampering. Currently, automotive vehicles do not have safety devices capable of informing whether the fuel is adulterated or not, and this analysis is carried out by regulatory bodies using instruments. This work develops a method for real-time and non-invasive determination of ethanol adulteration. An Arduino-based system acquires the vibration signal emitted by the engine of a running vehicle, and the signal is then processed using a technique for extracting chaotic signal characteristics called density peaks. Experimental tests were carried out in a combustion engine for the development and validation of the proposed method. No work was found in the scientific literature that proposes to estimate fuel with the car running, as the technique proposed in this work, and this is the first time that chaos theory is applied to the identification of adulterated fuel. The long-term autocorrelation function was observed for vibration signals obtained with the engine running from fuel with different percentages of ethanol, which revealed a long-term decay that evidences the chaotic behavior of the signal. A linear trend was demonstrated between the chaotic variable SAC-DM and the percentage of ethanol in the mixture with a correlation coefficient of 0.96.