| Popis: |
Despite the degree of railway electrification in many EU countries is higher than 50%, the diesel-driven railway vehicles continue to play an important role. As known, internal combustion engines, especially diesel engines, have also long been recognized as a significant source of pollutant emissions contributing to poor air quality, negative human health impacts and climate change. The future emissions regulatory control programs and the fuel-saving requirements for the new diesel engines for railways applications push worldwide OEMs, suppliers and scientific communities to investigate more advanced and alternative propulsion systems in which the diesel engines could still play an important role. Thus, the design of new power trains becomes more challenging considering the even more strict emission and efficiency targets. In this context, numerical simulation represents an essential tool in the entire development and optimization process of power trains. This study focuses on the numerical assessment of three different models of the same engine, characterized by different model accuracy, in order to evaluate the trade-off between model accuracy and computational time. The evaluation is carried out by performing the new emission standard Non-Road Transient Cycle (NRTC) applying the EU Non-Road Mobile Machinery (NRMM) directive to rail diesel vehicles. This work considers a 560 kW Heavy-Duty (HD) diesel engine. Regarding the models, the second and the third model are derived from the first one through an appropriate numerical procedure. The first, more accurate 1D model, is adequate when a deeper system analysis is required (i.e. wave dynamics, turbo-matching, etc.), while, for the evaluation of the global performance, the simplest model approach is more appropriate for complex systems, such as a hybrid powertrain. Indeed, the simplest model, despite its lower accuracy, shows good predictive results in terms of cumulative fuel consumption and cumulative NOx emissions over a transient homologation cycle. Moreover, for the lowest model accuracy, the real time factor is significantly lower compared to the more detailed one of about 250 times. |
