Predictive Engineering Uses Autodesk to Analyze Airflow of Respirator Mask

When the National Institute of Science and Technology (NIST) needed to come up with a safer, more effective respirator mask for firefighters, it turned to a company that has made its living combining finite element analysis (FEA) with computational fluid dynamics (CFD): Predictive Engineering Inc. in Portland, Oregon.

Led by CEO and owner George Laird, Predictive     Engineering has executed some 750 projects over the last six years, working with clients’ engineering teams to deliver the best possible structural and thermal/fluid designs.

Respirator masks have historically been designed with a lot of silicone rubber that can cover a range of faces. 

But, like Sansabelt slacks, the one-size-fits-all philosophy only goes so far.  When face profiles differ too much from the norm, respirators can suffer from air leakage or a contamination threat.

Pictured on the left: The respirator mask used in the NIST study conducted by Predictive Engineering.

NIST is conducting a multi-year program to improve the safety and effectiveness of full-face respirator masks. The goal is to combine free and easy airflow with optimum protection that doesn’t cause discomfort.

Predictive Engineering started by conducting an analysis of the fitting process between the respirator mask and a human head.  Respirator seal and head geometry were provided to Predictive Engineering as IGES data generated from 3D laser scanning.

FEA software was used to parse the skins together and create a clean manifold skin.  The composite skin was then used to generate a quad-dominate mesh for the respirator and a smooth tet mesh for the head.

The completed model was brought into another FEA program for fit and contact analysis. During the simulation, the mask was pulled against the face and allowed to seal. Analysis was then done to show seal pressures along the contours of the face.

Extensions were added at the inlets and outlets
of the mask for Autodesk CFdesign analysis. 
This was done to allow the airflow to develop
before it has to bend around the mouth.

 

Autodesk Simulation CFD software was used to assess airflow for the mask. This was quite tricky since the original geometry was not representative of the flow passage within the respirator. Predictive Engineering added extensions at the inlets and outlets of the mask model to better simulate how air flows before it has to bend around the mouth.

 

The inhalation model analysis in Autodesk Simulation CFD at
resting respiration. The convergence traces at the bottom of
the image show a stable transient analysis run with little variance
during the solution.  The graphic above shows the flow lines 
at the end of the solution and at near zero flow.

 

CFD analyses were conducted at resting and working flow rates.  Resting was based on a volumetric flow rate of 7.5 liters per minute while working was represented by a flow rate of 40 liters per minute.  Inhalation and exhalation were analyzed separately for both flow rates.

“It was impressive how accurately Autodesk Simulation CFD handled the transient flow conditions for inhalation and exhalation through the use of ramped flow-rate curves,” says Laird.  “When the CFD results were checked for convergence, the mass balance error was less than one percent.”

 

Both FEA and CFD simulations led to significant findings.

From its FE analysis, Predictive Engineering determined that the modeling of human skin is best represented as a flexible bag of viscous fluid, not as a semi-elastic solid as has been done in previous research outside of NIST. 

CFD analysis indicated that air flow within the respirator mask is not optimized and could be improved with some minor geometric changes.

These and other findings are scheduled for publication under the NIST banner with a co-authorship to Predictive Engineering for meeting project goals on time and on target.