Descriptive and inferential statistics as an exhaust emission comparative tool between different engine operating conditions and fuels. Application to highly oxidized biodiesel blended with primary alcohols

Abstract

Despite the great electrification that vehicle fleet is expecting in the coming years, internal combustion engines still play an important role in the transport sector. New regulations for polluting emission reduction and economy decarbonization, together with lower availability of fuels from non-renewable sources, have led to the search of renewable non-polluting fuels. This is a fundamental research sector in the next decades. Biodiesel, with a long history, has proven to be a good candidate. However, diversification will be the key to success in terms of sustainable mobility, avoiding negative dependency in the future. In this context, discarded oil is key to both produce an economically affordable biodiesel and to valorize this residue. This research studies the possibility of using a binary mix of fuels, including primary alcohols (propanol and pentanol), at 10 and 20% v/v, with highly-oxidized dumped oil biodiesel. These mixtures have been tested in a diesel engine running at 1300, 1700 and 2400 rpm, under 25% and 46% engine load for each rotational speed. To check the significance and reproducibility of obtained data, polluting emissions and engine input parameters have been evaluated through descriptive and inferential statistics. Results indicate that 20% 1-propanol/biodiesel blend allows mitigating the increase in fuel consumption of biodiesel compared to fossil diesel fuel. Moreover, overall soluble organic fraction (SOF) and unburned hydrocarbon emissions are lower for biodiesel and its blends than for diesel fuel. Furthermore, the combination of the use of 10% propanol and maximum EGR valve opening can achieve a significant reduction of NOx emissions. As a final conclusion, the proposed statistical study has shown to be a reliable tool to compare emissions and input parameters from different tests, especially under intra-urban conditions. This overcomes the large data variability from emission tests.