RESEARCH ARTICLE

Copyright © 2015 American Scientific Publishers Journal of Medical Imaging and

All rights reserved
Printed in the United States of America
Health Informatics
Vol. 5, 1–5, 2015

Numerical Simulation of Contaminant Control in

Multi-Patient Intensive Care Unit of Hospital
Using Computational Fluid Dynamics
Tikendra Nath Verma∗ and Shobha Lata Sinha
National Institute of Technology Raipur, Department of Mechanical Engineering, Raipur 492010, Chhattisgarh, India

The complex hospital environment requires special attention to ensure healthy indoor air quality (IAQ). It is
necessary to protect the patient and healthcare workers against hospital acquired pathogens/infections. A CFD
analysis supported by measurement data has been carried out to simulate the temperature distribution, air
flow pattern in the ICU and the contaminant dispersion from the patient. Numerical model solves conservation
of mass, momentum and energy equations. The contaminated particle (infection) which is simulated with a
Lagrangian particle tracking model and the same has been carried out by simulation. CFD analysis is used to
simulate and compare the removal of contaminated particle using a number of different ventilation systems. The
study concerns about the health risk of the airborne diseases (bacteria, fungus, viruses etc.) from the patient
to the other people in the ICU of hospital. A healthy environment can be achieved by minimizing the risk of
contamination through appropriate filtration and air distribution scheme. It is observed that stagnant zone near
the doctor and patient is not healthy. Therefore suitable ventilation arrangement and position of patient (bed)
must be provided for healthy environment in the ICU.

Keywords: Numerical Simulation, ICU, Contaminant Particle, Lagrangian Model.

1. INTRODUCTION the exhaled air from speaking was around 49 and 78 respec-
In India, hospitals are normally designed for multi-patient in a tively. Sun and Wang4 have presented a CFD-based virtual test
room. Economically, these rooms are the most efficient in which method for control and optimization of indoor environment by
patients can be placed along one corridor or in two rows, facil- combining a ventilated room with a ventilation control system.
itating the supervision of patients and reducing the amount of The ventilated room and its dynamic ventilation control system
time which nurses spend in travelling. Some pollutants of spe- are represented by a computational fluid dynamics (CFD) model
cial concern to indoor air quality are tobacco smoke, formalde- and models of the temperature sensor, PID controller and actuator
hyde, VOCs, radon, ozone and aerosols etc.1 In addition, odours, and VAV damper model respectively. Niu et al.5 have investigated
carbon dioxide and moisture which have significant effects on the transmission of respiratory droplets between two seated occu-
indoor air quality, particularly in densely occupied spaces. It is pants equipped with one type of personalized ventilation (PV)
necessary to identify those contaminants and their sources within device using round movable panel (RMP) in an office room.
the building as well as to establish acceptable concentrations in The office was ventilated by three different total volume (TV)
indoor air, since more than 8000 chemical species have been ventilation strategies, i.e., mixing ventilation (MV), displacement
identified in the indoor environment.2 Kwon et al.3 have ana- ventilation (DV), and under-floor air distribution (UFAD) sys-
lyzed the initial velocity and the angle of the exhaled airflow tem. Concentrations of particles with aerodynamic diameters of
from coughing and speaking of 17 males and 9 females using 0.8
m, 5
m, and 16
m as well as tracer gas were numer-
Particle Image Velocimetry (PIV) and acrylic indoor chamber. ically studied in the Eulerian frame. Yau et al.6 have focused
The results showed that the average initial coughing velocity on the ventilation of multiple-bed hospital wards in the trop-
was 15.3 m/s for the males and 10.6 m/s for the females, while ical climate, taking into account the design, indoor conditions
the average initial speaking velocity was 4.07 m/s and 2.31 m/s and engineering controls. The required indoor conditions such as
respectively. The angle of the exhaled air from coughing was temperature, humidity, air movements and indoor air quality in
around 38 for the males and 32 for the females, while that of the ward spaces are summarized based on the current guidelines
and practices. Tripathi and Moulic7 have analyzed the ceiling
∗
Author to whom correspondence should be addressed. type diffusers for the combinations of inlet streams on outlet

J. Med. Imaging Health Inf. Vol. 5, No. 5, 2015 2156-7018/2015/5/001/005 doi:10.1166/jmihi.2015.1500 1

RESEARCH ARTICLE J. Med. Imaging Health Inf. 5, 1–5, 2015

Table I. Expressions of
,  , S for various conservation equations for Table III. Temperature on different walls (in  C).
laminar flow.
East West North South Ceiling
Equation Continuity u-momentum v-momentum w -momentum Energy
31 25 17 28 48

1 U v w T
Note: Walls Temperature have been selected from ISHRAE handbook-2007, for Raipur
1 1 1 1
 0 (C.G.) Region, during summer season.
Re Re Re RePr
−dp −dp Gr −dp
S 0 + T 0
dx dy Re2 dz

locations including the variation of angles at inlet. The airflow

analysis indicates that for a single ceiling air diffuser, it is advan-
tageous to adopt a variation in the inclination angle at inlet to
achieve maximum air distribution profiles. The combinations of
inlet and outlet position reduce the short-circuiting of supply air,
providing additional uniform temperatures and room airflow in
the occupied zone. It is a genuine need to rethink the existing
ventilation design of intensive care unit of hospital to improve
pathogens/infection control and capabilities.

1.1. Governing Equation

The governing equations can be written in the General Transport Fig. 1. Room layout of multi-patient ICU (case 1–3).
Equation format and given as follows:


   • Patient is considered to be lying on the bed in the east-west
=  − ui  + S
t xi xi direction;
• When one patient is considered to be infected, other patients
The diffusion coefficient  and the source term S for various are assumed to produce no contamination;
equations are given in Table I. • The movement of only one particle has been considered during
Boussinesq’s approximation has been used to simplify the study for clarity in the figures;
buoyancy term. The equations are discretized into algebraic equa- • Contaminants have been considered mass less.
tions by the finite volume method (FVM) and Semi-Implicit
Pressure Linked Equation (SIMPLE) algorithm is adopted. The
Boussinesq model is employed to consider the buoyancy effect.8 1.2. Layout of Intensive Care Unit
Table II gives the details of variables, and for various conserva- Figure 1 shows the Multi-Patient intensive care unit room of size
tion equations. Table III gives the different temperature on walls. 4 m×3 m×6 m. The room consists of a four patients, one doctor,
The numerical simulation is based on following assumptions: bed, light and enclosure. Inlet opening and a door are assumed
• The bed, doctor, stool, patient and enclosure are of the shape to be situated on south wall whereas outlet is on north wall. The
of rectangular boxes; inlet and outlet area is kept constant throughout the simulation as
• The walls, floor and ceiling of the room are considered to be 0.6 m × 0.4 m. Table IV shows the dimensions of items available
well insulated; in ICU.

Table II. Notations for governing equations in cartesian co-ordinates for turbulent flow.

Equation   S

Continuity 1 0 0




1  p  u  v  w
u-momentum u + t − +  +  + 
Re Re x x x y x z x





1  p  u  v  w Gr
v-momentum v + t − +  +  +  + T
Re Re y x y y y z y Re2




1  p  u  v  w
w -momentum w + t − +  +  + 
Re Re z x  z y  z z  z
1 t
Energy T + 0
ReP r ReP r t
1 t
Turbulence kinetic energy k + Gk −  − Gb
Re k Re
1 t 
Heat dissipation rate  + C f G − C2 f2  + C3 Gb

Re  Re k 11 k
1 t
Local mean age  + 0
ReSc ReSc t

Notes: Where P r = 07; P r t = 08; Sc = 1; Sct = 07.

2
J. Med. Imaging Health Inf. 5, 1–5, 2015 RESEARCH ARTICLE

Table IV. Dimension of items available in the ICU.

Item/Occupancy Bed Doctor Door Patient Stool

Dimensions (in mm) 800 × 800 × 1700 300 × 1700 × 500 700 × 1800 300 × 1500 × 500 500 × 500 × 400

Table V. Supply air flow rate (AR) and air change rate (ACH). of supply and exhaust opening i.e., 0.6 m × 0.4 m. In all nine
Inlet Supply Supply and
cases, level of inlet position is constant i.e., 2300 mm above the
velocity Reynolds no. air flow (SR) ACH exhaust opening floor and outlet position is also constant i.e., 1700 mm from the
S.no. (m/s) (Re) (m3 /s) (/hr.) (m2) west wall. The distances of outlet from the floor are 300 mm,
1 05 15911 0.289 6 0.6 × 0.4 1400 mm, 2300 mm respectively.
2 075 23867 0.433 9
3 1 31823 0.578 12
2. RESULTS AND DISCUSSION
2.1. Case 1–3: Inlet 600 mm in Front of West WALL
(V = 0
5, 0.75, and1.0 m/s)
In this case, inlet is located 600 mm in front of the west wall and
2300 mm above the floor. Outlet is located 300 mm above the
floor and 1700 mm in front of west wall as shown in Figure 1.
The inlet velocity considered is 0.5, 0.75 and 1.0 m/sec, air
change rate per hour (ACH) of 6, 9 and 12. Inlet temperature is
20  C.
Figures 3(a)–(c) show velocity vector plots on three planes i.e.,
x = 10 m, 2.0 m and 3.0 m respectively on which the flow pat-
tern changes drastically. In Figure 3(a), it is clearly observed that
clean cold air is entering through inlet, moves horizontally and
bends near the north wall and falls down before leaving through
the outlet. The cold main stream moves above the occupied zone
of patients as well as doctor. One clockwise recirculation cell is
formed between the second patient and north wall and another
small anticlockwise recirculation zone is formed in the top right
Fig. 2. Planes for plotting velocity vector. corner of the plane. Figure 3(b) shows the velocity vector of ICU
on plane x = 2.0 m (mid-plane of ICU). One anticlockwise recir-
The inlet velocities considered are 0.5, 0.75, 1.0 m/sec, air culation cell is formed in the left side of the plane. The cold air
change rate per hour (ACH) of 6, 9, 12 and inlet temperature is leaving the ICU through outlet can be easily observed clearly.
20  C. Figure 3(c) shows the velocity vector of ICU on plane x = 3.0 m.
Table V shows different inlet velocity, Reynolds Number, Sup- It is clearly seen from the velocity vector that almost all the air is
ply air flow Rate (SR), air change rate (ACH), and dimension traveling from north wall to south wall. One small anticlockwise

(a) VV on plane x = 1.0 m (b) VV on plane x = 2.0 m

(c) VV on plane x = 3.0 m

Fig. 3. Velocity vector (VV) on different plane with-out enclosure (WOE).

3
RESEARCH ARTICLE J. Med. Imaging Health Inf. 5, 1–5, 2015

(a) Contamination through Patient 1 (b) Contamination through Patient 2

(c) Contamination through Patient 3 (d) Contamination through Patient 4

Fig. 4. Movements of contaminants coming from the mouth of different patients in ICU (WOE) (inlet velocity = 0.5 m/s).

recirculating zone is formed between patient 3 and 4. Another

small anticlockwise recirculating zone is formed between patient
4 and south wall.
Figures 4(a)–(d) shows that the movement of massless con-
taminated particle from the mouth of patient. Figure 4(a) shows
that the contaminated particle starts from the mouth of patient
1 and it simply leaves through the outlet without affecting the
doctor/nurse/other patients. The total time taken by the mass less
contaminated particle is approximately 3.2 minutes to leave the
room of intensive care unit (ICU) of hospital.

Table VI. Effect of contaminant on ICU environment and time required

by it to leave the ICU.

Case no. 1 2 3 4 5 6 7 8 9
Time required Patient 1 32 29 918 183 966 84 931 703 2.01
to leave Patient 2 102 113 42 68 47 398 481 24 8.92
the contaminants Patient 3 27 25 41 54 475 713 86 42 3.44
for Patient 4 47 238 195 485 213 218 385 216 7.46
Fig. 5. Room layout of multi-patient ICU (case 4–6).

4
J. Med. Imaging Health Inf. 5, 1–5, 2015 RESEARCH ARTICLE

The inlet velocity considered is 0.5, 0.75 and 1.0 m/sec, air
change rate per hour (ACH) of 6, 9 and 12. Inlet temperature is
20  C. In this case, the airflow pattern and movement of contam-
inant particle is same as previous case except the time required
by contaminant particle to leave the room which is mentioned in
Table VI.

2.3. Case 7–9: Inlet 2800 mm in Front of West Wall

(V = 0.5, 0.75, and 1.0 m/s)
In this case, inlet is located 2800 mm in front of the west wall
and 2300 mm above the floor. Outlet is located 300 mm above
the floor and 1700 mm in front of west wall as shown in Figure 6.
The inlet velocity considered is 0.5, 0.75 and 1.0 m/sec, air
change rate per hour (ACH) of 6, 9 and 12. Inlet temperature is
20  C. In this case, the airflow pattern and movement of contam-
inant particle is same as previous case except the time required
by contaminant particle to leave the room which is mentioned in
Table VI.
Fig. 6. Room layout of multi-patient ICU (case 7–9).

3. CONCLUSION
Figure 4(b) shows that the contaminated particle starts from The studies have been carried out at room inlet velocities of 0.5,
the mouth of patient 2 and moves through tortuous path. The 0.75 and 1.0 m/sec in ICU using k- model. Table VI shows
path of contaminated particle is relatively above the occupied the effect of contaminant on doctor / patients and time required
zone and leaves through the outlet without affecting the doctor or by it to leave the ICU room. It is observed that ventilation pro-
other patients. The total time taken by the mass less contaminated vided in Case 8 is the most appropriate as the contaminants take
particle is approximately 10.2 minutes lo leave the room. In this 2.4 minutes to leave the room, whereas in Case 04, the contam-
case, particle moves for larger time in comparison to the case inant particle takes longer time, approximately 48.5 minutes to
of patient 1 (Fig. 4(a)). Figure 4(c) shows that the contaminated leave the ICU. Therefore ventilation provided in Case 04 is not
particle starts from the mouth of patient 3 and then it moves in suitable for hospital.
the direction of south wall and subsequently it exits through the
outlet. The overall time taken by the mass less contaminated par-
ticle is approximately 2.7 minutes to leave the room of intensive References and Notes
care unit of hospital. It does not affect any of the patients as 1. H. B. Awbi, Ventilation of Buildings, 2nd edn., Spon Press, Taylor and Francis
Group, London and New York (2003).
well as doctor. Figure 4(d) shows that the contaminated particle
2. H. D. Goodfellow, Keynote address, ventilation-past, present and future, Pro-
starts from the mouth of patient 4, then it moves towards south ceedings of 2nd International Symposium, on Ventilation for Contaminant Con-
wall and leaves through the outlet without affecting the doctor trol, London (1988), pp. 5–11.
or other patients. The total time taken by the mass less contam- 3. S. B. Kwon, J. Park, J. Jang, Y. Cho, D. S. Park, C. Kim, G. N. Bae, and
A. Jang, Study on the initial velocity distribution of exhaled air from coughing
inated particle is approximately 4.7 minutes to leave the room and speaking. Chemosphere 87, 1260 (2012).
of ICU. 4. Z. Sun and S. Wang, A CFD-based test method for control of indoor environ-
By observing the movement of contaminated particle, it is ment and space ventilation. Building and Environment 45, 1441 (2010).
5. Q. He, J. Niu, N. Gao, T. Zhu, and J. Wu, CFD study of exhaled droplet trans-
advisable to keep the nurse-station and health care staff near the mission between occupants under different ventilation strategies in a typical
inlet to keep them healthy. office room. Building and Environment 46, 397 (2011).
6. Y. H. Yau, D Chandrasegaran, and A. Badarudin, The ventilation of multiple-
bed hospital wards in the tropics: A review. Building and Environment 46, 1125
2.2. Case 4–6: Inlet 1700 mm in Front of West Wall (2011).
(V = 0.5, 0.75, and 1.0 m/s) 7. B. Tripathi and S. G. Moulic, Investigation of air drafting pattern obtained from
In this case, inlet is located 1700 mm in front of the west wall the variation in outlet positions inside a closed area. Journal of Applied Fluid
Mechanics 5, 1 (2012).
and 2300 mm above the floor. Outlet is located 300 mm above 8. B. E. Launder and D. B. Spalding, The numerical computation of turbulent
the floor and 1700 mm in front of west wall as shown in Figure 5. flows. Computer Method in Applied Mechanics and Engineering 3, 269 (1974).

Received: 1 September 2014. Revised/Accepted: 31 March 2015.