A Novel 1D Co-Simulation Framework for the Prediction of Tailpipe Emissions Under Different IC Engine Operating Conditions 2019-24-0147
The prediction of the pollutants emitted by internal combustion engines during driving cycles has been a challenge since the introduction of the emission regulation legislation. During the last decade, along with the more tightening limits and increased public concern about the matter of air quality, the possibility of simulating various driving tests with cost effective computing facilities has become a key feature for modern simulation codes. Many 1D simulation tools are available on the market, offering real time models capable of achieving the simulation of any driving cycle in limited time frames. These approaches are based on the extreme simplification of the engine geometry and on the adoption of engine maps, which, for any engine operating condition, give the engine output in terms of power, or torque, and of exhaust gas composition. Specific fluid dynamic models are used to track the composition along the exhaust system and, with the aid of ad-hoc modules, to evaluate the conversion efficiency of after-treatment devices, such as TWC, GPF, DPF, DOC, SCR and so on. This work is based on the idea of joining two different 1D simulation tools in a co-simulation environment, realizing a strict coupling between the two codes. The main idea is to allow a precise 1D simulation of the unsteady wave action and gas motion inside the intake and exhaust system, without resorting to oversimplified geometrical discretization, and to relying on advanced combustion models for the calculation of in-cylinder emissions. The simulation of the after-treatment systems is then performed resorting to a steady state model with detailed chemistry models. In this scenario, the choice of the coupling strategy is a key issue, since an unsteady model is coupled to a steady-state one. In particular, this last element may flatten the unsteady pattern of the flow, imposing unreal boundary conditions at the engine exhaust manifold and increasing the risk of having misleading results. For this reason, the chemical and the fluid-dynamic behaviour have been decoupled, letting the 1D solver to propagate pressure waves along the complete system, preserving all the characteristic lengths (acoustic and fluid dynamic). For every aftertreatment device a specific steady-state 1D model is used to model the chemistry and the heat transfer process occurring inside these devices. The boundary conditions for the steady state solver are given by the 1D unsteady model, suitably averaged over the time. The steady-state solution is then sent back to the 1D unsteady solver and suitably over imposed: temperature of the substrate, the friction coefficient and gas composition. This co-simulation environment is validated against a real engine configuration which was measured at EMPA labs. A 4 cylinders, turbocharged, CNG engine is simulated at different loads and revolution speeds and the results are compared with the simulations to verify the prediction of engine performance and pollutant emissions.