Centralized exhaust system simulation and parameter optimization

Mingming Lu, Miao Qian, Haoyang Zhang, Guangwei Sun, Mingyue Yang, Yao Zhao, Kai Zhang, Chong Wang, An‐ran Wang, Tongfu Zheng

Abstract

With the increase in high-rise buildings in cities, public flue exhaust systems have a direct impact on residential air quality and the living environment. Although existing studies have analyzed the problems in public flue exhaust systems through computational fluid dynamics (CFD) numerical simulations and experimental tests, these studies often lack an in-depth exploration of the specific impacts of individual components in the system. To solve this problem, this study not only thoroughly analyzes the key components in the public flue system, such as branch pipes, check valves, and tee pipes, but also develops a parametric public flue simulation system software based on C# and verifies the accuracy of the simulation through experiments. The study first determines the key parameters affecting the comprehensive resistance coefficient of the branch pipe, check valve, tee pipe, and other components through CFD simulation and experimental testing. Subsequently, a visualization program is developed using the C# language, which can quickly simulate and visualize the flow changes in the flue according to different building parameters such as the number of floors, height of floors, and size of the flue. The results confirm that the program can efficiently predict the flow distribution under different design options, providing a practical tool for the optimal design and performance evaluation of public flues, which is important for improving the air quality of the living environment.

Introduction

In modern high-rise residential architecture, kitchen smoke is managed through a centralized "public flue"—a vertical shaft shared by multiple apartments. Improper design of these systems leads to "smoke backflow," where cooking odors from one unit infiltrate others. While engineers have traditionally used Computational Fluid Dynamics (CFD) to study these systems, most models are macro-level and rely on fixed, simplified resistance values for individual components.

This research addresses the micro-level details of the system, specifically investigating how the geometries of branch pipes, check valves, and tee joints affect overall performance. By moving away from "one-size-fits-all" constants, the study aims to create a parametric model that adapts to varying building heights and floor plans.

Schematic diagram of the centralized exhaust system components including the main flue, branch pipes, and check valves.

Methodology

The researchers employed a multi-faceted approach to develop and validate their simulation tool:

  1. Numerical Simulation (CFD): Using the standard $k-\epsilon$ turbulence model in Fluent, the team simulated airflow at critical confluence points where branch pipes merge with the main vertical flue.
  2. Mathematical Modeling: Governing equations were derived from aerodynamics theory to account for pressure losses along the main flue ($P_y$) and at confluence points ($P_{1-3}$ and $P_{2-3}$).
  3. Software Development: A custom visualization program was built using C#. This tool allows designers to input specific building parameters—such as floor count, floor height, and flue dimensions—to instantly predict flow distribution without the computational cost of full CFD suites.
  4. Experimental Validation: The mathematical models were tested against physical test benches, including a 33-layer vertical flue (Vanke Tower) and a 6-layer horizontal flue, measuring actual pressures and flow rates.
Flowchart of the iterative calculation process used in the parametric simulation software to determine pressure values across different floors.

Key Insights and Results

The simulation model demonstrated high accuracy when compared to real-world physical tests, achieving a Pearson correlation coefficient of 0.9768 for full 33-layer system simulations.

  • The "Bottom Floor" Penalty: The study quantified a significant disparity in exhaust capacity. At a 100% start-up rate (all hoods active), a bottom-floor unit may only move 185 $m^3/h$, while a top-floor unit moves 648 $m^3/h$. This is due to the massive cumulative pressure resistance faced by lower levels.
  • Component Resistance Factors: Research revealed that branch pipe resistance is relatively independent of air volume and length (averaging ~2.76). Conversely, check valve resistance is almost entirely dependent on the opening angle ($ heta$), following an exponential relationship: $\xi = k_1 e^{k_2(90-\theta)}$.
  • Design Recommendations: To mitigate the performance gap on lower floors, the study suggests that developers utilize higher-performance fans specifically for lower-level apartments to compensate for the system's inherent pressure resistance.

Limitations and Future Work

While the current model is highly accurate for new installations, it assumes clean internal surfaces. Real-world factors such as grease accumulation and oil scale buildup will eventually change the surface roughness ($K$) and narrow the effective pipe diameter. Future iterations of the C# software will aim to integrate these "aging" factors to provide long-term maintenance and performance predictions for high-rise building management.

Summary updated

Abstract

With the increase in high-rise buildings in cities, public flue exhaust systems have a direct impact on residential air quality and the living environment. Although existing studies have analyzed the problems in public flue exhaust systems through computational fluid dynamics (CFD) numerical simulations and experimental tests, these studies often lack an in-depth exploration of the specific impacts of individual components in the system. To solve this problem, this study not only thoroughly analyzes the key components in the public flue system, such as branch pipes, check valves, and tee pipes, but also develops a parametric public flue simulation system software based on C# and verifies the accuracy of the simulation through experiments. The study first determines the key parameters affecting the comprehensive resistance coefficient of the branch pipe, check valve, tee pipe, and other components through CFD simulation and experimental testing. Subsequently, a visualization program is developed using the C# language, which can quickly simulate and visualize the flow changes in the flue according to different building parameters such as the number of floors, height of floors, and size of the flue. The results confirm that the program can efficiently predict the flow distribution under different design options, providing a practical tool for the optimal design and performance evaluation of public flues, which is important for improving the air quality of the living environment.
BibTeX
@inproceedings{lu2024centralized,
  author = {Mingming Lu and Miao Qian and Haoyang Zhang and Guangwei Sun and Mingyue Yang and Yao Zhao and Kai Zhang and Chong Wang and An‐ran Wang and Tongfu Zheng},
  title = {{Centralized exhaust system simulation and parameter optimization}},
  booktitle = {Scientific},
  year = {2024},
  doi = {10.1038/s41598-024-72160-1},
}
Figure