Modelling of multiple biodiesel-emitted nitrogen oxides using ANN approach.

• Properties of Multiple biodiesel blends tested as per ASTM standards. • Mean gas temperature increases with increase in biodiesel concentration. • 18.93% increment in NOx is due to predominant effect of mean gas temperature. • ANN feed-forward backpropagation (TRAINLM and LEARNGDM) model predicts NOx by 0.99. This research paper focuses on modelling of nitrogen oxides emitted by diesel engine for multiple biodiesel blends. A lot of research work has observed that the properties of biodiesel blends affect the nitrogen oxide emissions. To this aim, total of fifteen blends of multiple biodiesels are prepared on vol. % by using four non-edible category biodiesels. The suitability and quality of the biodiesel and diesel blends are tested through a characterization and found within the permissible limits of ASTM. The experimentation has been carried out on a direct injection compression ignition (DICI) engine with constant speed and varying loading condition from no load to full load in a step of 25%. During the experimentation, the NOx emissions are measured using a Netel MGA-2 exhaust gas analyzer. The study revealed that several properties such as viscosity, density, mean gas temperature, etc. affect NOx emission. In addition, NOx emission increases with an increase in BPs. The artificial neural network (ANN) is performed by considering physicochemical and thermal properties as a function. The ANN predicts the estimated NOx with an accuracy of 0.99. [ABSTRACT FROM AUTHOR]