Ethanol Content Estimation in a Flex-Fuel Motercycle using Engine Speed Measurements

The performance of engine management system of flexible fuel vehicles relies on the estimation of the ethanol content in the fuel blend. Many previously proposed estimation approaches involve additional redundant sensors which increase production costs. In this research work, we proposed a simple ethanol estimation technique which uses only the measurement of engine-speed at a fixed starting condition. The engine speed response was modelled as a first-order system. It was found that the steady-state characteristics of the model were able to capture the effect of changing fuel composition. By applying a nearest-mean classifier-based scheme, we could select the best feature from the model. Five different blends, including E10, E20, E50, E85, and E100, were used to formulate the model and experimentally validated the proposed technique.