Inverse design of an indoor environment using a filter-based topology method with experimental verification

Area of Science:

• Building Science
• Environmental Engineering
• Computational Fluid Dynamics

Background:

• Optimizing indoor environments for health, comfort, and energy efficiency is crucial.
• Computational fluid dynamics (CFD) and topology optimization offer potential for designing optimal air supply systems.
• Existing CFD-based adjoint methods with area-constrained topology can result in scattered inlet distributions.

Purpose of the Study:

• To introduce and verify a filter method for smoothing intermediate results in CFD-based inverse design.
• To optimize the number, size, location, and shape of air supply inlets and their parameters.
• To achieve a healthy, comfortable, productive, and energy-efficient indoor environment.

Main Methods:

• Developed a filter method to enhance CFD-based adjoint topology optimization for air inlet design.
• Applied the method to a 3D, non-isothermal, asymmetrical office model with pre-set inlets.
• Validated the optimized design by simulating and experimentally measuring indoor environmental parameters.

Main Results:

• The filter-based topology method successfully smoothed inlet distribution compared to previous methods.
• The optimized design achieved desired levels of thermal comfort, health, productivity, and energy efficiency.
• Experimental validation in an environmental chamber closely matched CFD simulation results for air temperature, velocity, and mean age of air.

Conclusions:

• The proposed filter method is effective for inverse design of air supply inlets.
• This approach enables the creation of superior indoor environments.
• The findings have practical implications for HVAC system design and building performance optimization.