At the rear of the vehicle an end acoustic silencer is attached to the exhaust system. This is primarily to reduce noise emissions for the benefit of passengers and bystanders. Due to the location of the end acoustic silencer conventional thermal protection methods (heat shields) through experimental means can not only be difficult to incorporate but also can be an inefficient and costly experience. Hence simulation methods may improve the development process by introducing methods of optimization in early phase vehicle design.
A previous publication (Part 1) described a methodology of improving the surface temperatures prediction of general exhaust configurations. It was found in this initial study that simulation results for silencer configurations exhibited significant discrepancies in comparison to experimental data. This was mainly due to the inability to represent complex fluid flows through the components of the silencer, which was greatly simplified in the simulation models and software utilised. The following investigation addresses the errors noted in the previous publication and considers a further improvement to the end acoustic silencer models.
The investigation covers the combination of alternative heat transfer models and sophisticated geometrical modeling. It was found that pipe friction must be considered to calculate the pressure losses within the silencer systems. The bending influence on internal flow conditions and resulting heat transfer effects were considered through the utilization of convective augmentation factors. The perforation through the internal pipe walls were taken into account by increasing the frictional factors within the Nusselt number models. Finally the absorption material within the chambers of the silencer system was modeled utilizing a multi-layer arrangement within the simulation software.
Two vehicle models were considered, the first being a 4 cylinder engine and the second a 6 cylinder engine. Both models simulated two operating points, 35 km/h hill climb and 250 km/h. The simulation was then validated through experimental climatic wind tunnel tests, where temperature probes were attached to highly thermal sensitive regions.
Utilizing the proposed methodology, the simulation yielded results that featured strong correlations to experimental data. Within the majority of probed locations a temperature difference of less than 50 K was achieved. Future work consists of improving the areas of discrepancy, particularly surrounding chambers of a hollow type.