Acoustic analysis of engine exhaust ducts: a comprehensive empirical/1D simulation approach
Idrees, Muhammad Danish (2023-02-22)
Idrees, Muhammad Danish
22.02.2023
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe20231123148581
https://urn.fi/URN:NBN:fi-fe20231123148581
Tiivistelmä
Internal combustion engines, renowned for their instantaneous and high-energy output, are in high demand in the heavy-duty sector, especially within the marine industry. However, they pose significant challenges with atmospheric and noise emissions. Excessive noise impacts human health and has adverse effects on the harbor environment due to the significant noise emissions produced by vessels. One of the main causes that deliberately contributes to this noise is the exhaust stack. Studying how noise behaves in these exhaust ducts can help improve their design and reduce these noise emissions.
This thesis aims to study the behavior of low-frequency acoustic in exhaust ducts through experiments and simulations, while ignoring thermal and mean-flow effects. The study is based on two experiments: one using the main exhaust pipe of the W4L20 medium-speed engine, and the other using a plastic pipe (3.2 m), called the reference experiment. The thesis focuses mainly on the reference experiment. This reference experiment was modeled using the 1D modeling software GT-SUITE. The aim of this study is to identify the parameters that affect the accuracy of noise measurements with a perspective to use them to calibrate airflow simulation models. To this end, a unique approach is adopted to define the acoustic source in the GT model using input data obtained from experiment. In addition, a comparative analysis between the GT model and experimental results has been conducted to identify discrepancies. The thesis also includes a fundamental analysis of the experimental results for the main exhaust pipe. The emphasis is on understanding the phenomenon of acoustic propagation through complex airflow paths.
Experimental results showed that the fundamental mode of wave propagation remained dominant. In the reference experiment, this mode succeeded in exciting the resonance frequencies of the plastic pipe. However, this mode failed to excite the resonance of the exhaust pipe. During noise propagation, it was observed that the effect of pipe bends on the SPL was negligible. Furthermore, simulation results based on the GT model showed that the reliability of these results depends on the accuracy of the input data measured during the experiment. Measurements taken outside the source opening are very sensitive to sensor placement. They showed 8 to 10 dB difference in SPL at a small distance of only 9 mm. It was concluded that for duct-based simulations, input data should be measured from inside the source pipe. The parametric study using the GT-model showed that the resonant frequencies decrease with increasing pipe length. However, the increase in pipe diameter and pipe bend had no significant effect on the frequency spectrum.
This thesis aims to study the behavior of low-frequency acoustic in exhaust ducts through experiments and simulations, while ignoring thermal and mean-flow effects. The study is based on two experiments: one using the main exhaust pipe of the W4L20 medium-speed engine, and the other using a plastic pipe (3.2 m), called the reference experiment. The thesis focuses mainly on the reference experiment. This reference experiment was modeled using the 1D modeling software GT-SUITE. The aim of this study is to identify the parameters that affect the accuracy of noise measurements with a perspective to use them to calibrate airflow simulation models. To this end, a unique approach is adopted to define the acoustic source in the GT model using input data obtained from experiment. In addition, a comparative analysis between the GT model and experimental results has been conducted to identify discrepancies. The thesis also includes a fundamental analysis of the experimental results for the main exhaust pipe. The emphasis is on understanding the phenomenon of acoustic propagation through complex airflow paths.
Experimental results showed that the fundamental mode of wave propagation remained dominant. In the reference experiment, this mode succeeded in exciting the resonance frequencies of the plastic pipe. However, this mode failed to excite the resonance of the exhaust pipe. During noise propagation, it was observed that the effect of pipe bends on the SPL was negligible. Furthermore, simulation results based on the GT model showed that the reliability of these results depends on the accuracy of the input data measured during the experiment. Measurements taken outside the source opening are very sensitive to sensor placement. They showed 8 to 10 dB difference in SPL at a small distance of only 9 mm. It was concluded that for duct-based simulations, input data should be measured from inside the source pipe. The parametric study using the GT-model showed that the resonant frequencies decrease with increasing pipe length. However, the increase in pipe diameter and pipe bend had no significant effect on the frequency spectrum.