TL;DR
This study uses advanced computational fluid dynamics (CFD) and acoustic modeling to analyze how coke buildup in ethylene cracking furnace tubes affects flow-induced noise. The findings reveal that coke deposition significantly amplifies noise, primarily from dipole sources concentrated in the tube's "necking" region, offering a new way to monitor furnace health and improve safety.
What Problem Does This Paper Address?
Ethylene cracking furnaces are crucial for polymer production, but a common issue is coke deposition on the inner walls of the furnace tubes. This buildup not only impacts process efficiency but also poses safety risks. While predicting coke deposition is challenging, understanding its effect on the furnace's acoustic behavior could provide an early warning system. This paper addresses the need to characterize the flow-induced noise generated by these furnace tubes, specifically before and after coke formation, to better understand and potentially predict coking-related issues.
Methods
The researchers employed a hybrid numerical approach combining Large Eddy Simulation (LES) for fluid dynamics and a generalized Lighthill's acoustic analogy model for acoustics.
- Large Eddy Simulation (LES): This technique was used to simulate the unsteady flow field within the furnace tubes. LES is known for its accuracy in capturing turbulent flow structures, which are the primary drivers of flow-induced noise. The study used the Smagorinsky-Lilly model to determine subgrid-scale turbulent viscosity.
- Generalized Lighthill's Acoustic Analogy: This model, an extension of Lighthill's original theory, was used to predict the acoustic field. It accounts for the presence of solid boundaries and moving media, making it suitable for this application.
- Hybrid Approach: The CFD results (velocity, pressure, density) were mapped to an acoustic grid using the iCFD module in Actran. Acoustic predictions were then made through interpolation between CFD and finite element grids. Finally, the time-domain acoustic sources were analyzed using discrete Fourier transform (DFT) and a direct frequency response method.
Datasets
Not discussed in the paper. The study relies on numerical simulations rather than experimental datasets for training. However, the results were validated against experimental data from a previous study (Lv & Ji, 2011) for flow noise prediction.
Key Insights and Ideas
- Coke as a Noise Amplifier: Coke deposition dramatically increases flow-induced noise. The study quantifies this, showing significant growth in sound power and turbulence intensity as coke thickness increases.
- Dipole Sources Dominate: The primary noise sources in coked tubes are identified as dipole acoustic sources. These are strongly correlated with turbulence intensity near the tube wall.
- "Necking" is the Hotspot: The "necking" region of the furnace tube, where the diameter constricts, is identified as the primary location for acoustic source generation. This is likely due to increased turbulence and pressure pulsations in this area.
- Frequency Signatures: The noise energy is predominantly concentrated in the medium-to-low frequency range (300-1500 Hz). Specific peak frequencies (e.g., 794.56 Hz, 631.65 Hz, 517.32 Hz) emerge with increasing coke thickness, potentially serving as diagnostic indicators.
- OASPL as a Coking Indicator: The Overall Sound Pressure Level (OASPL) curve along the central axis of the furnace tube shows distinct patterns. The presence of a local peak in OASPL can be used to classify coking levels as "mild" or "severe," offering a practical method for monitoring.
Key Results
- Sound Power Level Increase: As coke layer thickness increased from 5 mm to 15 mm, the sound power level increased by 60.5%.
- Turbulence Intensity Increase: Concurrently, turbulence intensity increased by 58.5%.
- OASPL Values: For coke layer thicknesses of 10 mm and 15 mm, the OASPL values reached 79.25 dB and 119.08 dB, respectively, at the "necking" region.
- Noise Emission: The noise levels observed, particularly at higher coking degrees, significantly exceed typical noise emission limits, highlighting a potential hazard.
- Radiation Pattern: The noise radiation pattern exhibits characteristics of dipole acoustic sources, with distinct directional lobes.
Limitations and Future Work
- Idealized Coke Layer: The study simplifies the coke layer as smooth and uniform, disregarding surface irregularities. Real-world coke deposits are often rougher, which could influence acoustic behavior.
- Simplified Geometry: While the study focuses on a radiation coil section, real furnace systems are more complex.
- Experimental Validation: While validation was performed against a prior study, direct experimental validation of the noise characteristics for various coking levels in actual furnace tubes would strengthen the findings.
Future work could involve more detailed modeling of irregular coke surfaces, extending the analysis to more complex furnace geometries, and conducting extensive experimental validation to refine the predictive capabilities of this acoustic monitoring approach. The authors suggest that these findings can contribute to enhanced safety measures and more effective coking management strategies in ethylene cracking furnaces.
Summary updated