Predicting indoor particle dispersion under dynamic ventilation modes with high-order Markov chain model

Area of Science:

• Environmental Engineering
• Computational Fluid Dynamics (CFD)
• Indoor Air Quality (IAQ) Modeling

Background:

• Mechanical and natural ventilation are crucial for maintaining indoor air quality (IAQ) by removing airborne contaminants.
• First-order Markov chain models are used for IAQ simulations but struggle with particles exhibiting significant inertia.
• Accurate simulation of particle behavior is essential for effective IAQ management.

Purpose of the Study:

• To develop a novel weight-factor-based high-order (second- and third-order) Markov chain model for indoor airborne particle simulation.
• To simulate particle dispersion and deposition under both fixed and dynamic ventilation conditions.
• To improve the accuracy of IAQ models for transient airflow scenarios.

Main Methods:

• Utilized computational fluid dynamics (CFD) to solve flow fields under various ventilation modes.
• Developed and implemented second- and third-order Markov chain models incorporating weight factors.
• Validated the first-order Markov chain model against literature simulation and experimental data.
• Tested different weight factor combinations to optimize model performance.

Main Results:

• The proposed second-order Markov chain model accurately predicts particle dispersion and deposition under fixed and consecutively changed ventilation modes.
• High-order models offer improved accuracy compared to the traditional first-order model with minimal increase in computational cost.
• Optimal weight factors were identified for second-order (λ1=0.7, λ2=0.3, λ3=0) and third-order (λ1=0.8, λ2=0.1, λ3=0.1) models for reduced prediction errors.

Conclusions:

• The developed high-order Markov chain model provides a more accurate alternative for simulating indoor airborne pollutants, including particles and gases.
• The model is capable of handling transient ventilation modes, offering flexibility in IAQ analysis.
• Further improvements in state transfer matrix construction and CFD data processing will enhance the model's applicability for fast IAQ predictions.