Tese: An optimized method for automotive performance predictions using different mixtures of ethanol and gasoline
Aluno(a) : Leonardo Pedreira PereiraOrientador(a): Sergio Braga e Antonio Villela
Área de Concentração: Termociências
Data: 12/08/2021
Link para tese/dissertação: https://doi.org/10.17771/PUCRio.acad.56809
Resumo: Vehicle acceleration performance is an important feature to be evaluated when internal combustion engines and new fuels are being developed. Predicting this parameter is also of great significance, once automotive track tests requires long periods of time to be done and high costs with equipment, rental of the track, hiring people and displacement of vehicles and fuels. In addition, the results are directly affected by track surface irregularities and variations in weather conditions such as ambient pressure, temperature, air humidity, wind speed and crosswinds. Therefore, this work aims to use the data collected in bench tests with an internal combustion engine, in order to computational modelling speed recovery tests of its respective car. Doing these experiments in laboratory has the advantage of a greater control of the room ambient conditions and the engine operating parameters. Also, it reduces the expenses related to on-track testing. The proposed methodology simulates the traction force on the wheels based on the measured torque in dynamometer or from the pressure curves inside the combustion chamber with the aid of friction models for spark ignition engines. In order to validate the proposed model, it became necessary to perform speed recovery tests with the car on a chassis dynamometer. Also, seven different mixtures of ethanol and gasoline were used, and it was concluded that pure anhydrous ethanol promoted a higher acceleration capacity in most of the experiments but it had higher fuel consumption. It was also seen that hydrated fuels reduced performance. The simulations demonstrated a high precision in relation to the experiment, with a recovery time difference average of 0.51 seconds and standard deviation of 0.078. Also, the acceleration performances had errors smaller than 5.25%.