Thermo-fluid Dynamics Analysis and Air Quality for Different Ventilation Patterns in An Operating Theatre

At present there is a great deal of research on ventilation system efficiency for providing the most effective airflow pattern for contaminant removal (microorganisms i.e. fuWngi, bacteria, viruses; chemical i.e. waste anesthetic gases, CO₂ and particulate matter) and assuring the control of surgical site infection risk. Computational fluid dynamics (CFD) modelling and simulation, compared with experimental monitoring results and literature evidences, are widely applied for providing useful indications on proper indoor microclimate conditions and IAQ [1]. Usually, three-dimensional time-dependent CFD models are developed to assess the airflow patterns, air velocity and temperature distribution due to laminar unidirectional (vertical downward) air flows [2], in other cases to turbulent air flows and also to the air curtains so as to achieve IAQ and optimal air asepsis [3].

Some authors have used CFD for a three dimensional analysis of thermal comfort and contaminant removal inside hospitals: checking model validity by experimental literature evidence and data comparison, they calculated the predicted mean vote (PMV), contaminant removal effectiveness (CRE) and mean contaminant concentration in the breathing zone by transient simulation results, in order to assess HVAC system

efficiency. They investigated the effect of horizontal location of supply and exhaust grilles also for maintaining adequate comfort conditions for patients and surgical staff [4]. Many literature has used CFD simulation to assess colony forming unit (cfu) dispersion and evaluate air flow distribution effectiveness in real OTs [. In a more specific research the influence of periodic bending movement of the medical staff combined with bacteria carrying particle distribution on the airflow field and temperature distribution have been investigated [5].

Most of these studies consider correct use, on-design operational conditions of OTs. Some new analyses have been proposed concerning moving object effects on the airflow inside OTs under correct and incorrect use conditions [6] [7]. In the most recent literature, the assessment of ventilation efficiency in OTs, including air flow distribution under different configurations and different positions of diffuser supply velocity, is studied by applying CFD simulation and full-scale experimental investigations [8]. A recent study based on experimental measurements and CFD simulation shows the efficacy of different ventilation systems in removing bacteria and particles using adducted return strategy to induce air flow from clean to less clean zones [9]. In the literature there is not much that addresses the analysis of the efficiency of a ventilation system in an OT, equipped with ante-chamber and/or air locking system for guaranteeing air pressure difference, indoor microclimatic conditions and IAQ, minimizing the air flow inlet from non-clean zones to the OT. A recent research work of the present authors has dealt with the efficiency of air curtain application for producing a vertical downward air curtain around the laminar diffuser using experimental data and CFD modelling of the studied real OT [10]. Actually, most of the work on this subject deals with different applications [11].

Linear slot diffusers have been used to create a vertical downward air curtain around the ceiling unidirectional flow diffuser towards low-level exhaust grilles, to avoid the in-ward flow under the unidirectional diffuser. The authors demonstrated that a clear correlation between the unidirectional flow and air curtain flow rates cannot be assumed and this cannot be read in a better indoor microclimate and IAQ conditions [12].

In the present work we investigated which scheme of the ventilation system inside the OT (i.e. dimensions and position of the inlet and outlet air diffusers) is effective to minimize risk of poor IAQ and surgical site infections caused by real operational use conditions. From a well-known starting phase, concerning our previous study of a real OT under operating use conditions [7], different configurations of the present ventilation scheme were studied with transient CDF simulations and corresponding results were compared and discussed calculating some crucial ventilation efficiency and thermal comfort parameters.

Simulations were carried out by commercial software allowing multi-physical approach through solutions of the related governing equations by a Finite Element method [16] Solutions for the variables, air velocity (U) air pressure (p) , temperature (T ), carbon-dioxide concentration in air (CO₂), mean age of air (τ ) and CFU concentration in air (CFU) were carried out. The basic formulation of the partial differential equation formulation (PDE) used for computations is:

$\frac{\partial(\rho \phi)}{\partial t}+\nabla \cdot(\rho \phi U)=\nabla \cdot(\Gamma \nabla \phi)+\Lambda$   (1)

Where ρ stands for fluid density, U velocity vector, Γ diffusion coefficient and Λ the source term. For fluid-dynamical analysis a RANS approach considering the fluid Newtonian and the flow incompressible was used. A standard two-equation (k-ε) closure scheme was applied for turbulence solution [17] [18].

1.png

Figure 1. Geometry of the studied OT: dashed boxes indicate location and type of incoming/outflow section for each configuration

Table 2 provides all the different symbols used in the generic eq.(1) applied for each governing equation. Values of physical properties used for materials and objects are listed in Table 3. Independently from the specific ventilation scheme, at the inlet sections for air incoming into the system we applied as boundary conditions, a constant velocity for the OT (Table 1) and assumed 0.66 m/s for the corridor diffusers (6 ACH). Turbulence intensity at inlet was set equal to 5%. Values of the other first-type boundary conditions applied at the inlet sections are reported in Table 4. For breathing modelling, a sinusoidal trend for inhaled/expired air and for CO₂ emission rate into the room was assumed. In steady state simulations, the corresponding root mean square (RMS) values was used as magnitude for air velocity and CO₂ mass flow rate at the nostrils of the medical staff, patient and persons standing in the corridor. Table 5 provides the set of parameters used to simulate people presence, in terms of breathing, heat production, CO₂ and CFU emission in the indoor environment. Outflow conditions were considered for all the dependent variables at the recovery grids.

Table 1. Ventilation air flow rate for each ventilation scheme

Table 2. Formulation of variables and parameters in Eq. (1)

Table 3. Physical properties of materials in numerical models.

Table 4. Values of variables at the inlet boundaries.

Table 5. Parameters for human presence simulation.

At each solid/fluid interface, logarithmic wall functions were applied to the near wall airflow, that was considered parallel to the wall and being in a wall offset equal to one hundred viscous units. Turbulent production was assumed to equal dissipation at walls. For the remaining dependent variables impermeable/insulation conditions at solid walls were assumed.

A non-uniform and non-structured computational mesh made up of the second order tetrahedral elements was applied. Steady solutions of discrete equations were carried-out by applying an iterative dumped Newton-Raphson scheme [19] based on the discretized PDE linearization by a first-order Taylor expansion. Algebraic systems of equations coming from differential operator discretization were solved by a PARDISO package, a direct solver particularly efficient for solving unsymmetrical sparse matrixes by a LU decomposition method. The convergence criterion was set to 1E-5. Time integration of governing equations for transient simulations was performed by applying an Implicit Differential-Algebraic (IDA) solver  which  uses  the  variable

order and variable-step-size of Backward Differentiation Formulas (BDF). Because the time-marching scheme is implicit, a nonlinear system of equations must be solved at each time step. Newton algorithm was exploited to solve this nonlinear system of equations. All computations were carried-out on a workstation with two 64-bit 6-core/12-thread processors speeding up to 2.3 GHz of frequency and handling 128 GB of RAM.

Starting from simulation results on air temperature and velocity fields, CO₂ and particle concentration expressed as CFU/m³ were investigated and some IAQ indexes, with regard to contaminant removal and infection risk control, were calculated. These indexes refer to the average values of dependent variables in the different zones of the OT (i.e. BZ, OZ, PZ and TV). At the beginning for the IAQ assessment, the mean age of air (τ), that quantifies the average lifetime of air at a particular location of the room for a steady airflow achieved was calculated as well as a dependent variable, as explained in the modelling section. Referring to Fig 5, a measure of air freshness due to different ventilation schemes can be deduced. At the inlet opening it was assumed to be zero (100% fresh).

5.png

Figure 5. Mean age of air on a horizontal slide (z = 1.4 m)

The average value of τ was computed in the different zones and called as τ_Z, where Z _j means the generic j-zone. Comparison of these values with the theoretical residence time of air inside the OT (defined as the ratio between the total volume of the room (VTV, m³) and the mass flow rate of incoming ventilating air (V_vent, m³/s)), shows that the ratio is always higher than 1. This comparison consists Air Change Efficiency (ACE) index evaluation. This index, which measures how effectively ventilation systems replace the air in a room with fresh air, is expressed as:

$A C E=\frac{\frac{V_{T V}}{V_{v e n t}}}{\tau_{Z j}} \cdot 100$     (2)

The Ventilation Effectiveness (VE) index that measures how quickly a contaminant is removed from an air volume by quantifying the efficiency with which the internal pollutant is diluted or removed was calculated. It depends on the air flow patterns, concentration field computed for CO₂ and particles, and can be expressed as:

$V E=\frac{C_{E}-C_{S}}{C_{Z J}-C_{S}}$     (3)

CE is the mean value of contaminant concentration (i.e. CO₂ and particles) calculated at the air-recovery grilles (Exhaust), CS is the contaminant concentration at the air inlet diffusers (Supply) and C_ZJ is the mean value of the contaminant concentration in a specific zone. Similarly, the Contaminant Removal Effectiveness (CRE) index representing the ratio between the concentration of contaminants at the exhaust point and the mean value of contaminant concentration within a specific zone, was calculated as:

$C R E=\frac{C_{E}}{C_{z J}}$      (4)

VE and CRE indexes were computed using CO₂ and particle concentration results. Particle concentration value was assumed to be zero at the inlet air diffusers (CS=0) then, VE and CRE expressions correspond to each other.  Referring to ACE index, the V_2 produces the results shown in Table 6. This ventilation models does not include the vertical air curtain located at the sliding door, but has the vertical air curtain  located at the ceiling supply diffuser along only two sides, corresponding to the underlying sides of the bed. VE and CRE indexes compared to the mean age of air for the BZ and OZ, show better performances and efficacy of the modified ventilation system of V_1 and also for V_2. In particular, V_2 is characterized by a unidirectional horizontal air flow. This is the reason why the V_2 model provides the higher mean air age values for the BZ and OZ. Analysing PZ and TV, the situations inverts (Table 6): V_2 model provides similar results of the mean air age and the lowest ones among all the models. Diffusion Performance Index (ADPI) according to [20] was calculated. The ASHRAE Handbook of Fundamentals provides the basic definition of ADPI which statistically relates the space conditions of temperature and air velocity to occupant thermal comfort, underlining that high ADPI values are desirable as they represent a high comfort level mainly due to increased probability of good ventilation air mixing. Considering a chosen number of discrete points in the OT (146 locations for 1 < x < 5; 5 < y < 10 and 0.5 < z < 2.5) ADPI index was computed as the percentage of test points that meets the performance parameters provided by the combination between the Effective Draft Temperature (−1.5 K < EDT < 1.0 K) and air velocity (U < 0.35 m/s). The EDT is defined as

$E D T=\left(T-T_{s e t}\right)-8 \cdot(U-0.15)$     (5)

Table 6. The IAQ indexes for the OT zones and for each ventilation schemeWhen the effective draft temperature is between −1.5 K and +1.0 K and the air velocity less than 0.35 m/s, a high percentage of people feel comfortable. Results comparison shows that V_2 model guarantees the best thermal comfort conditions, because their results involve a lower comfort conditions for all the considered OT zones. V_2 gives also the higher ADPI index value. The ACE index results for V_1reache the highest values and even now comparable with that of V_2. Looking at the CFU concentration results for each OT considered zones, the V_1, model provides the lower values, but V_2 ventilation scheme from this point of view, is the best for BZ, OZ, PZ and then TV.

The authors thank Prof. Pietro Vitali (Hospital Hygiene Unit, University Hospital of Parma), Prof. Cesira Pasquarella (Department of Biomedical, Biotechnological and Translational Sciences, University of Parma), Prof. Roberto Albertini (Department of Clinical and Experimental Medicine, Unit of Medical Immunology, University Hospital of Parma) for their important support to the present research.