Personal ventilation to control airborne infectious diseases in … · To control infection spread,...

Eindhoven University of Technology

MASTER

Personal ventilation to control airborne infectious diseases in hospital patient roomscross-contamination of airborne infections tested by static and dynamical experiments with apersonal ventilation pillow

van der Sanden, N.P.M.

Award date:2012

Link to publication

DisclaimerThis document contains a student thesis (bachelor's or master's), as authored by a student at Eindhoven University of Technology. Studenttheses are made available in the TU/e repository upon obtaining the required degree. The grade received is not published on the documentas presented in the repository. The required complexity or quality of research of student theses may vary by program, and the requiredminimum study period may vary in duration.

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

https://research.tue.nl/en/studentthesis/personal-ventilation-to-control-airborne-infectious-diseases-in-hospital-patient-rooms(66dc8cc6-bbc8-42a9-a22d-8f0452bd12b9).html

Personal venti lation to control ai rborne infectious d iseases in hospital patient rooms

Eindhoven University of Technology July 2012 0

Personal ventilation to control airborne infectious diseases in hospital patient rooms Cross-contamination of airborne infections tested by static and dynamical experiments with a personal ventilation pillow.

July 2012 N.P.M van der Sanden Student number: 0571268 MSc Student Building Services Unit of Building Physics and Services Faculty of Architecture, Building and Planning Eindhoven University of Technology Prof. Dr. H.S.M. Kort First supervisor, Eindhoven University of Technology Dr. Ir. M.G.L.C. Loomans Second supervisor, Eindhoven University of Technology Dr. Ir. F. Franchimon External supervisor, BAM techniek



Preface This is the final report for the graduation project of Nard van der Sanden, for the Master of Science degree in Building Services at Eindhoven University of Technology. This project was initiated by the request of BAM techniek to find the possibilities for using personal ventilation for infection control in hospital patient rooms The report consists of a paper on the research which was performed and some related appendices. The paper gives a brief introduction of the subject, the method used for research, results and discussion of the research and finally a conclusion. In appendix A, a more detailed introduction is given, Appendix B-E give more information about the methodology, and Appendix G-I consist tables and graphs, which provide a more detailed view on the results section. I would like to take the opportunity to thank the people who supported me in this presentation. Firstly I would like to thank, Helianthe Kort, Marcel Loomans en Francesco Franchimon for their valuable input, both on the subject and on the process. They helped me greatly in introducing me in the subject of hospital ventilation and infection control, brighten up my approach and goals to get this project most of the time smooth working and finally critically support me during the experiments. I also would like to thank Wout van Bommel, who was always able to help me and think with me how to set-up the best possible experiments. Besides that I would like to thank my fellow students, colleagues at BAM techniek and family for providing an optimal work atmosphere.



Contents Preface ..................................................................................................................................................... 1

1. Abstract ...................................................................................................................................................... 3

2. Introduction ................................................................................................................................................ 4

2.1 Personal ventilation systems in hospital patient rooms .................................................................... 4

2.2 Dynamical actions ............................................................................................................................ 6

2.3 Aims of this study ............................................................................................................................ 7

3. Methodology .............................................................................................................................................. 8

3.1 Static and continuous movements with CO2 tracer gas. ................................................................. 10

3.2 Single movement measurements .................................................................................................... 11

3.3 Measurement parameters ............................................................................................................... 12

4. Results ...................................................................................................................................................... 13

4.1 Static experiments .......................................................................................................................... 13

4.2 Continuous movements .................................................................................................................. 14

4.2.1 Influence of movements ...................................................................................................... 15

4.3 Single time movements .................................................................................................................. 16

5. Discussion ................................................................................................................................................ 18

5.1 Method ........................................................................................................................................... 18

5.2 Static situation ................................................................................................................................ 20

5.3 Continuous Movements ................................................................................................................. 22

5.4 Single time movements .................................................................................................................. 23

6. Conclusion ............................................................................................................................................... 23

7. Further research ....................................................................................................................................... 24

8. Acknowledgements .................................................................................................................................. 24

References ......................................................................................................................................................... 25

Appendix A: Introduction, Nosocomial infections ..................................................................................... 28

Appendix B: Pictures measurement equipment ........................................................................................... 34

Appendix C: Measurement equipment......................................................................................................... 36

Appendix D: Measurement protocol. ........................................................................................................... 37

Appendix E: Calibration .............................................................................................................................. 40

Appendix F: Conversation hospital ventilation ........................................................................................... 45

Appendix G: Results; performance of PV when only one pillow is active .................................................. 47

Appendix H: Results graphs ......................................................................................................................... 49

Appendix I: Statistical analysis .................................................................................................................. 84

Appendix J: Used Matlab M-files ............................................................................................................... 90



1. Abstract Introduction: Hospitals face big threats with the spread of an infectious disease among patients. The contribution of spread through the air of these infections is seen to be more and more important. Personal ventilation is a new technique in hospital patient rooms, which provides fresh air directly to the patient’s breathing zone to avoid cross contamination of a non-infected person by an infectious person in the same room. Personal ventilation has shown to be very efficient in the protection of people from cross infection. However frequently occurring movements in hospital patient rooms may influence the air flow pattern in a room and by that movements may influence the spread of infectious particles in a room. Material and methods: The influence of movements on contaminant transport in a hospital patient room ventilated with a personal ventilation pillow is examined in this study. Tracer gas and particle spread are done in a full-scale experimental two-person test room, where static experiments, experiments with a continuous walking person and experiments with single time movements as a single-time moving person, turning over bed-sheets and opening a door are investigated. Results: Effectiveness of 95-98% at breathing level of the receiver patient is reached in the reduction of contaminant concentration by use a personal ventilation pillow in static situation. During continuous movements PV is still able to reduce contaminant concentration with 91% at the breathing zone of the receiver patient. Movements do only cause significant differences in contaminant concentration when the movement is nearby the investigated zone. Conclusion: From the results of the full-scale experiments it can be concluded that the personal ventilation pillow works effective on contaminant reduction on the receiver patient in a mixed ventilated room as a high effectiveness of the personal ventilation pillow of 98% in the static situation is achieved. During continuous movements personal ventilation is still able to reduce contaminant concentration with 91% at the breathing zone of the receiver patient. It can be concluded that personal ventilation systems perform better in contaminant reduction at the receiver patient’s breathing zone in both static, continuous dynamical situations as well as single time movement situations. Discussion and further research: Personal ventilation with lower room ventilation rates may also perform well on cross-contamination and this seems interesting for the building industry as a ventilation configuration with lower and better directed airflow rates is less energy consuming and therefore decreases the exploitation costs. However this is not tested in this study and effects of lower room ventilation rate on both patients as healthcare workers should be investigated in further research. More research is also needed on the subject of comfort of patients lying on a breathable pillow. Keywords: Personal ventilation, hospital patient room, infection control, contaminant transport, cross-contamination, dynamical actions



2. Introduction There has been strong and sufficient evidence to demonstrate the association between ventilation and the control of airflow directions in buildings and the transmission and spread of infectious diseases such as measles, tuberculosis, chickenpox, anthrax, influenza, smallpox, and severe acute respiratory syndrome (SARS) (Li et al., 2007). Hospitals face even bigger threats with the emergence of new infectious diseases such as avian influenza (Yam et al., 2011). Also recent epidemics as the 2011 Klebsila pneumonia in Rotterdam’s Maasstad Hospital elucidate the danger of infection breakouts in hospitals. If particles carrying pathogens are inhaled by a susceptible individual and deposited in a suitable location in the respiratory tract, infectious disease may occur. Therefore, it is critical to investigate how the human exhaled droplets from infected human bodies transport or disperse indoors and how to control them to maintain a safe indoor environment (Chen & Zhao, 2010). Especially in hospital environments, where there is a high risk on airborne infectious diseases due to high emission rates and the higher susceptibility to diseases from hospital patients, investigation of the transportation path of infected particles is necessary. It is concluded that although contact-spread is the principle route of transmission for most infections, the contribution of airborne micro-organisms to the spread of infections is likely to be greater than is currently recognized (Beggs, 2003). The movement of particles in ventilated areas is influenced by many factors, such as airflow pattern, particle properties, geometry configurations, ventilation rates, supply and exhaust diffuser locations, internal partitions and thermal buoyancy due to heat generated by occupants and/or equipment etc. (Zhao et al., 2004). Several studies (Yam et al., 2011; Knibbs et al., 2011) agree that indoor ventilation with good air quality control minimizes the spread of airborne respiratory and other infections in hospitals. Infections which originate in the hospital environment on patients who didn’t suffer on infection symptoms by entering the hospital are called nosocomial infections. In the Netherlands 11% of patients (n=2661) get an nosocomial infection disease, during their hospital stay (Hopmans et al., 2007). Nowadays hospitals in the Netherlands are more and more forced by governments and insurance companies to lower their cost and in relation to this, decrease the hospitalization time of patients. Patients who developed nosocomial infections do have a prolonged length of stay and therefore increase the pressure on waiting lists which leads to decreasing patient satisfaction. Besides that nosocomial infection could even lead to mortality. Infection diseases could be harmful by an infective dose of 1 to 10.000 infected particles depending on the disease and the condition of the patient (Kowalski, 2002). See Appendix A. In the design process, evidence based design principles in order to meet up to government regulations and to decrease the hospitalization costs are very welcomed in both the building and the hospital environment

2.1 Personal ventilation systems in hospital patient rooms To control infection spread, a new technique in hospital patient room ventilation is introduced; personal ventilation (PV). Personal ventilation is usually based on jets of air directed to a person’s face (Melikov, 2004). Different designs are designed where PV systems should always be supplemented with a general ventilation system in the room. Melikov uses a system where the supplied air to the breathing zone is filtered by Ultraviolet germicidal irradiation (UVGI) to kill or render the infectious particles harmless by the use of ultraviolet energy. Nielsen et al. (Nielsen et al., 2008) designed a personalized ventilation system which utilizes the situations where the head or the body is in natural contact with surfaces as chairs, beds, pillows, mattresses, clothing etc. Those surfaces are designed also to be a supply opening of fresh air, for example by the use of fabric as a diffuser. An example where a breathable pillow is used is shown in Figure 1. Another design by Victor Cheung shows a retractable hood (Figure 2) which reduce the cross-infection risk by reducing the emission from the source patient with a local exhaust.



Figure 1: Air supply pillow intended for a hospital bed.

(Nielsen, 2009)

Figure 2: Retractable hood design. Idea of Victor Cheung, JRP Hong Kong. (Li et al., 2003)

2.1.1 Performance of PV systems As personal ventilation is a relatively new concept in hospitals, there are only a few studies known by the author of this article, testing the performance of these systems. In office situations personal ventilation is already generally used and several studies (Cermak et al., 2006; Pantelic et al., 2009; Melikov, 2004) show that personal ventilation will always be able to improve the inhaled air quality in rooms with mixing ventilation and indoor air quality in rooms with displacement ventilation with regard to pollution emitted from the floor and from human-produced contaminants. PV systems which supply clean air are able to decrease the inhaled pollutant concentration by factors between 2 and 50 times compared with total-volume ventilation alone. Pantelic concluded that PV is able to reduce the concentration of contaminated droplets in the inhalation zone and at a distance wide enough from the index patient, the momentum of cough jets could be bend away. In hospital patient rooms the local hood exhaust design is found to be very effective in removing the virus-laden aerosols. With the hood, the virus-laden aerosols originated from the patient's mouth can be captured fully even when the patient is more than 300 mm away from the exhaust (Li et al., 2003). Obviously, the patient's head should be covered under the hood to obtain the 100% capture efficiency. It is found that the exhaled airflow direction is also an important parameter. If the patient face is positioned outside the local exhaust and the exhaled airflow is directed to the surrounding, the virus-laden aerosols can escape into the test room. Other problems with such a device are that it is difficult to keep clean, it cannot be expanded the whole time and it is uncomfortable for the patient (Nielsen, 2009). A patient room ventilation systems with an air supply pillow obtains a personal exposure index εexp,PV higher than 10 for flow rates above 10 l/s (Nielsen, 2009). The consequence of this high personal exposure means that particles in the receiver manikin’s breathing zone can be reduced from a level of 15000 droplet nuclei m-3 with only downward ventilation, to a level of 1500 droplet nuclei m-3 when PV is supplemented. This exposure index is even increasing to 35 for a flow rate of 14 l/s. In an upward displacement ventilated room, the air supply PV system protects the patient in a bed, but it is also possible to use the system to reduce the emission from an infected patient (Nielsen et al., 2007a). The air supplied from the PV diffuser rises to the ceiling and can then be removed through the high located exhausts. This is important for the health care workers in the hospital, who cannot be protected by the PV system. A downward ventilation in a room, with a diffuser close to the PV device, decreases the effect of the PV system (Nielsen et al., 2008). When the patient room uses vertical downward ventilation from ceiling-mounted diffusers, and a high location of the exhaust openings, then the PV system is able to reduce the emission from a source patient and thus reduce the risk of airborne cross infection for other patients and health care workers in the patient room. High flow rates (�̇�pv > 12 to 20 l/s) from the PV pillow give the best result (Nielsen et al., 2008). Besides the improvement on minimizing cross infection, the PV system in patient rooms has of course all the features known from conventional office systems as e.g. the possibility to have individual control of the thermal environment (air temperature en humidity), which in itself could be a positive feature in a bed environment. Because of the infection control capability properties, PV may be used by high risk patients like oncologic patients, patients with a reduced immune system and internist patients, those



groups represent 20 % of patient room population (Trip & Bielleman, 2012). PV is especially efficient if the patients are bed-bound, and the effectiveness will be reduced when they are mobile (Nielsen, 2009).

2.2 Dynamical actions Air moves around a room and between rooms as a consequence of (Tang et al., 2011):

• Ventilation (either forced or naturally ventilated); • The movement of people, equipment, furniture and doors; • Buoyancy driven flows generated by heat of people and equipment; • The disturbances created by human respiratory activities, such as breathing talking and coughing.

Movement of air due to ventilation or buoyancy driven flows generally generates predictable air flows which are investigated in a reasonable number of studies as can be seen in the literature study. Moving objects usually disturb the air distribution and result in the fluctuations of velocity, pressure, temperature and concentration of contaminants within the room. Only a few studies have investigated this subject and unfortunately there is still no unique turbulence model, numerical scheme or numerical algorithm valid for solving all indoor air motions (Shih et al., 2007). Movement of persons Movement of people also changes the pressure, velocity and contamination concentration fields in a room and therefore plays a significant part in transporting infected air from one place to another. As the amount of movement increases, the concentration distribution in the room will move towards a fully mixed situation, however local phenomena are observed (Mazumdar et al., 2010). In contradiction to the fact that turbulence can improve mixing ventilation patterns, there are also potential dangers for patient rooms where turbulence can lead to material spreading to adjacent beds and cross contaminating other patients (Klettner et al., 2009). For typical turbulence values found in the hospital environment, a droplet of diameter 30 𝜇𝑚 at an initial height of 2m will be spread a distance of 2m due to turbulence caused by typical movements in an hospital room (Eames et al., 2009). Considering that these calculations have been carried out without a mean flow present, which is likely to be present in a hospital, they represent quite conservative estimates on droplet spread (Klettner et al., 2009). A moving person can carry a contaminant in the wake to positions far from the contaminant source (Mazumdar et al., 2010). It can be concluded that the faster the walking speed, the longer is the wake behind the human back (Wang & Chow, 2011). Experiments with two breathing manikins in a displacement ventilated room (Bjorn & Nielsen, 2002) showed that air exhaled through the mouth can be locked in a thermally stratified layer where contaminant concentrations several times the return concentrations can occur. Stratification of exhaled air will break down immediately when a moving subject is passing. The boundary flow around the body still offers some protection in these situations. However the protective effect of the boundary layer flow around the body of moving person will disappear already at a speed of 0.2 m/s. This means that a person inhales the same concentration as in the ambient air at breathing zone height. Although faster moving speeds result in larger velocity and pressure variations, those changes disappear and the flow field returns to the original state quickly, which usually takes less than 30 s after the moving person returns to the original position according to numerical measurements done by Shih (Shih et al., 2007). Door opening In hospital patient rooms, doors are many times opened and closed during a day and these actions cause airflow in the room. Door edge travels about 0.8m in about 2s, when it is opened relatively slowly (Tang et al., 2006), generating an airflow with speed of approximately 0.8/2 = 0.4 m/s. Numerical experiments done by Mazumdar (Mazumdar et al., 2010) shows that the concentration distributions changed with the processes of the door opening and closing. But as soon as the door was closed, the concentration quickly returned to that of the steady state within 10 to 15 s. As the door is often far away from the contaminant source, the impacts on contaminant concentration differences are not very large.



Bed sheets When sheets are moving up, contaminants are induced in the wake. By moving the sheets down over the patient, the contaminants induced were pushed out through both sides (Figure 3). The contaminant concentration returns to the initial level of the steady state in approximately 50 s (Mazumdar et al., 2010). In such cases dispersion does not occur from a single point in space as for a respiratory release, instead the dispersal of infected aerosols will vary in spatial location and intensity depending on the activity.

t=0s t=0.75s t=1.5s

Figure 3: Numerical investigation of moving bed-sheet with 1m/s up and immediately down (Mazumdar et al., 2010). Red implies a high concentration of contaminants, while dark-blue is a low concentration.

Another way of spread of infectious particles is the microbial shedding from our body which occurs all time. When a bacterium is shed into a textile fabric between the patient and the bed, either in his pyjama or directly on the sheet, the moisture and temperature in textile micro-environment promotes its proliferation (Borkow & Gabbay, 2008). Bed making generates dust and airborne microorganisms (Shiomori et al., 2001). The release into the air of particles contaminated with infectious aerosols such as MRSA may occur from the skin which is shed during routine activities such as walking and bed making (Hathway et al., 2011). Shiomori et al. found a 25 fold increase in the number of MRSA in the air immediately following bed-making. The bacteria levels in the air fell back to background levels within 30 minutes in a 37-bed otolaryngology–head and neck surgery unit. A numerical study done by Mazumdar et al. (Mazumdar et al., 2010) reveals that movements of walking people, moving bed sheets or opening a door may cause a swing in the contaminant concentration at the breathing level of sitting and standing positions for 10 to 90 seconds. The variation of the averaged contaminant concentration due to the moving objects was within 25% for all the cases studied. The closer the location of the moving object to the contaminant source, the larger was the change in the contaminant concentration. Since the variation only lasted for less than 90 s and the averaged contaminant concentration during the day did not change much, the variation would not likely change the average risk level in the patient room. Hence at locations close to the contaminant source, the risk of transmission is higher with the displacement ventilation system compared to the mixing ventilation system.

2.3 Aims of this study This study aimed at investigating the distribution of exhalation particles when a personal ventilation pillow is used in a hospital patient room. Personalized ventilation may reduce the risk of cross-infection in a downward ventilated room, and in some cases, it can also reduce the source of infection (Nielsen, 2009). However turbulence caused by dynamical actions plays an important role in the ventilation strategy of hospital rooms (I Eames, 2009). In the hospital environment there is much interest in infection control on an evidently based design basis. As movements in patient rooms may substantially influence the airflow pattern, the hospital building industry is interested in the performance of ventilation systems in combination with movements occurring in reality. These dynamic situations are rarely studied and there are no studies, which research the exposure to air contaminants in dynamical situations by use of a personal ventilation system in patient rooms. Therefore dynamic situations, for example, the entry of visitors or nurses in a patient room or the sudden turning over of sheets will be examined in this study. Main research question of this study is: Will personal ventilation systems in hospital patient rooms perform similar for infection control as non-personal systems in dynamical situations?



Therefore local air quality experiments are done in a mixed ventilated full-scale experimental two-person hospital patient room where situations are tested with no movements, a continuously moving person, a single time moving person, moving bed sheets and an opening door. Those situations are tested with PV for both patients, PV for one individually patient only. A situation with no PV is tested to compare the personal ventilation with commonly used mixed ventilation (Table 1).

3. Methodology Room used A full-scale experimental patient room with dimensions of 4.4m (long) x 3.6m (width) x 2.8m (high) is used for this study. This room has well insulated walls and outdoor weather conditions do not influence the measurements, as all the walls, ceiling and floor of the room are adjacent to the air conditioned indoor environment of the building physics laboratory at the Eindhoven University of Technology. Therefore the walls are considered to be adiabatic. The temperature outside the test room in the laboratory hall varies between 23 and 25 ℃ and the air supply temperature of the room ventilation system is 20 ℃, the designed room temperature is 24 ℃, which is in line with the guidelines of CIBSE and the American Institute of Architects (AIA, 2006; CIBSE et al., 2005). There is one door in the room and one window which is adjacent to a small blind room. In the middle of the room there are two fluorescent lightings which both have a load of 80 W each. The room ventilation system is of a mixing type, a ceiling mounted inlet diffuser of 30 cm x 30 cm is installed in the middle of the ceiling. The outlet with diameter 12 cm is located just above the door at a height of 2.15m above floor level. The total ventilation rate, which is the personal ventilation rate plus the normal room ventilation rate is 4 ACH, the same as used in comparable studies (Bjorn & Nielsen, 2002; Qian et al., 2006; Qian et al., 2008; Mazumdar et al., 2010). When personal ventilation was used during the experiments, the room ventilation rate is decreased with the flow rate of the PV system to ensure that the total ventilation rate remains unchanged. The exhaust air outlet outside the test room is situated 3.5m away from the air supply and exhausts in opposite direction to prevent for cross-flow.

�̇�𝑡𝑜𝑡𝑎𝑙 = �̇�𝑟𝑜𝑜𝑚 𝑣𝑒𝑛𝑡𝑖𝑙𝑎𝑡𝑖𝑜𝑛 𝑠𝑦𝑠𝑡𝑒𝑚 + �̇�𝑝𝑣 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 = 4 𝐴𝐶𝐻 [1] Personal ventilation system The personal ventilation system used during the experiments is a reproduction of the design of Aalborg University of Technology professor Peter. V. Nielsen (Nielsen et al., 2007b) and is based on a breathable pillow, where fresh air is supplied direct to the breathing area. The pillow has dimensions of 50 cm (length), 85 cm (width) and 5 cm (thickness). The pillow is constructed with an airtight bottom and sides, and a permeable textile top. The air permeability of the textile diffuser is 130 m3/m2/h at 80 Pa, which is tested with the FlowFinder combined with a pressure gauge. The air flow rate of the PV system is set to 15 l/s at which the system performs optimal according to Nielsen (Nielsen et al., 2007b), see Figure 4. The flow rate of the pillows is continuously monitored by a calibrated measuring tube. The fresh air is supplied to the pillow by a flexible tube of diameter 16cm, connected to the centre of the long side of the pillow (Figure 5). The temperature of the air which is supplied by the personal ventilation is isothermal to the ‘outside’ laboratorial temperature, which lies between 23 and 25 ℃.



Figure 4: Effectiveness of the breathable pillow with several air flow rates. (Nielsen et al., 2007b)

Figure 5: Fresh Air supply to the pillow

Manikins There are two manikins used in this study (Figure 6). The manikins are lying on beds, simulated by tables of dimensions, 200cm (length) x 100 cm (width) and 76 cm (height). The beds are situated with the head to the long wall of the room and there is 1 meter space in between the beds. Manikin 1, the source patient, is developed at the TU/e and is 1.65m tall with a black surface colour; he is dressed with an open jacket and jogging pants. His head has a circular diameter of 20 cm, the body has an oval diameter of 35 cm and the arms and legs have circular diameters of 15 cm. The temperatures of this manikin are controllable in ten body sections and the surface temperature of this manikin is 35 ℃ at his head, 34 ℃ at his body and 33 ℃ at his legs and arms. The total heat production of this manikin is 85 W, which represents the heat gain for a person in rest (CIBSE, 2006). This manikin has a breathing feasibility of constant flow of air of 6 l/s, the normal breathing volume of a person in rest (CIBSE, 2006). The position of the outflow of the breathing air is at the mouth of the source patient. The flow opening of 11 mm diameter means a velocity of 1 m/s, which is a normal breathing flow velocity (CIBSE, 2006). This volume of 6 l/min is divided in an outside air section of which its flow is continuously monitored by mechanical flow meters and in the tracer gas experiments this air is mixed with 100% CO2 gas of which the flow is controlled by a digital mass flow controller. The concentration of CO2 at dispersion at mouth level is 200000 ppm. Manikin 2, the receiver patient, is 1.80 m tall; is also made at the TU/e; and is divided in 3 sections. This manikin is metal-gray coloured and has no clothes. The head has got a diameter of 16 cm and a temperature of 35 ℃, the body has got a diameter of 36 cm and a surface temperature of 34 ℃. The legs have got a diameter of 16 cm and a surface temperature of 29 ℃. The total heat gain of this manikin is 85 W. This manikin has no breathing facility.

a) b) c) Figure 6: a): Overview of the experimental room. b): Source manikin with the breathing device at his mouth.

c): Receiver manikin



Measurement schedule The experiments are separated in three parts; section B where steady state experiments with continuous distribution of tracer gas are done; section C where those experiments will be repeated with addition of a continuous moving person in between the beds and sections D, E and F where single movement experiments are performed. As the CO2 sensors do not have a fast enough response time, this last part is done with continuous smoke gas distribution in combination with particle count sensors. See Table 1. A description of the whole measurement protocol is given in Appendix D Table 1: Measurement schedule (see also Appendix D) Test Orientation manikins Cases Movements Method Steady state case with tracer gas analyzing B1-B4 Face up 1. PV off for both patients 2. PV on for both patients

3. PV on at source patient only 4. PV on at receiver patient only

no CO2

B5-B6 Face to Face 1. PV off for both patients 2. PV on for both patients 3. PV on at source patient only 4. PV on at receiver patient only

no C02

Continuous moving person tracer gas analyzing C1-C4 Face up 1. PV off for both patients 2. PV on for both patients

3. PV on at source patient only 4. PV on at receiver patient only

Continuous moving person

CO2

C5-C6 Face to face 1. PV off for both patients 2. PV on for both patients 3. PV on at source patient only 4. PV on at receiver patient only

Continuous moving person

C02

Single time moving person with smoke gas and particle counts analyzing D1-D2 Face to face 1. PV off for both patients 2. PV on for both patients

single time moving person

Particles

Single time moving bed-sheets with smoke gas and particle counts analyzing E1-E2 Face to face 1. PV off for both patients 2. PV on for both patients

single time moving bed-sheet

Particles

Swinging door - with smoke gas and particle counts analyzing F1-F2 Face to face 1. PV off for both patients 2. PV on for both patients

single time moving bed-sheet

Particles

3.1 Static and continuous movements with CO2 tracer gas. A Bruel & Kjaer multi-gas monitor (type 1302) with a multi-point sampler (type 1303) was used to measure the CO2 concentrations in the room. The accuracy of the Bruel & Kjaer is 1.7 ppm. 5 samplers were used for measuring the concentration, see Figure 7; 1 is a sampler 50 cm above the source patient’s mouth; 2 and 3 are samplers in between the beds at a height of 1.2 and 1.65 m, which represent the position of the mouth of respectively a sitting and a standing health care worker/ visitor; position 4 is 50 cm above the receiver patient’s mouth and sampler 5 is at the position of the receiver patient’s mouth. This last position is on the top of the manikin’s head, when the position of the manikins is face-up and on the side of manikin’s head when the position is face to face. The Bruel & Kjaer uses 1 minute to sample and calculate the CO2 concentration for each sampling point and is not able to do this parallel for multiple points, so a whole measurement cycle takes 5 minutes. For each position there are 36-40 measurement values measured during 3-4 hours lasting measurement series

Figure 7: Position of the measurement points to measure CO2 concentration



The outside CO2 concentration, which is measured at the origin of the supply air flow tube and the exhaust CO2 concentration, measured inside the exhaust air flow tube are measured with Vaisala CaTec CO2 indicators. These sensors have an accuracy of 40 ppm. The supply air temperature is continuously monitored at the supply air flow rate, the room temperature is measured at point 3 (Figure 7) and the outside temperature, just outside the laboratory room. The case where a person is continuously moving through the room is simulated by a cylinder of height 1.9m and diameter 0.45 m moving over a rail (Figure 8). The cylinder moves for 3s over 2.4m then stops for 5 seconds, moves back also waits 5 seconds at the other end and this movement will continue the whole time. Measurement configuration and techniques are the same as in the static situation.

Figure 8b: Moving manikin

3.2 Single movement measurements Experiments with single movements are done to test the performance and robustness of the personal ventilation system under frequently occurring dynamical actions in hospital patient rooms. The first case tested is a walking person in between the beds (Figure 9a). This is represented by a cylinder of 1.9m high and 0.45m diameter travelling over a rail of 2.4m at 0.75 m/s and moving back after pausing at the head end for 10 seconds. Secondly moving bed-sheets at the source patient, travelling 0.75 m vertical and immediately move back at 1 m/s are simulated (Figure 10). The last case is an opening and immediately closing door at 1 m/s.

(a) (b) (c)

Figure 9: Single movements measurements: (a): A walking person in between the beds, (b): Moving bed-sheets at the source patient, (c): opening and closing of the room door

a) b) Figure 10: Moving bed-sheets. Bed sheets, a): Sheets are lying down on the source patient; b): Sheets are pulled up till upper position above the source patient.



As available CO2 sensors were not accurate and fast enough to measure the quick concentration differences during a single movement, other techniques are used here. There has been made use of smoke gas produced by a disco-smoke generator. This generator regular spreads smoke in 2.2 m3 buffer, where the smoke is mixed and variations in smoke densities are averaged out; as a results of this a constantly flow of smoke is transported to the breathing area of the source manikin. Due to sensor limitations it was not possible to measure the high particle concentration at the mouth of the source patient. To control the released particle concentration in the room, the concentration in the exhaust is measured, however especially larger particles could be suspended in the room or in the ventilation tubes. So the measured exhaust concentration does not give a direct relation to the released particles, but as the smoke generator is the main distributor of particles, the exhaust concentration can be used to see whether or not the concentration will be constant. As the most critical way of airborne spread is the spread from breathing zone to breathing zone, the smoke gas particles are only counted at the breathing area of the receiver manikin (Figure 7: point 5). Smoke particles are counted in the range 0.3-0.5 𝜇m and larger than 0.5 𝜇m, which is the size range of large viruses and small bacteria. Patients are only lying face to face, as results of continuous movement measurements gave comparable concentration levels for the face to face case as the face up case. Mazumdar (Mazumdar et al., 2010) concluded that air patterns returns to normal condition within maximum 90 s after a movement, so to be sure a double relaxation time of 3 minutes is taken, tests with 5 and 15 minutes relaxation time have been executed and gave the same results and so longer relaxation times do not have significant advantages. A total of 50-64 of movement experiments, divided in two comparable measurement series are done per case.

3.3 Measurement parameters Local air quality index The local air quality (LAQ) index is a measure of the local concentration of a contaminant in a point The local air quality index at the position p, 𝜀𝑝 is defined as 𝜀𝑝 = 𝐶𝑒𝑥ℎ

𝐶𝑝 [2]

Where,

cexh = concentration in the exhaust cp = concentration at point p

A value of: 1 will present a completely mixed situation. <1 Local concentration of a contaminant is worse than average concentration in the room. >1 Local concentration of a contaminant is better than average concentration in the room. The effectiveness of personalized ventilation The effectiveness of personalized ventilation is defined by Melikov (Melikov et al., 2002) as 𝜖𝑝,𝑝𝑣 = 𝑐𝑝,0−𝑐𝑝,𝑝𝑣

𝑐𝑝,0−𝑐𝑠𝑢𝑝,𝑝𝑣 [3]

The exhaust air CO2 concentration is the exhalation source of the source manikin csource, divided by the total room air exchange rate �̇�𝑡𝑜𝑡𝑎𝑙 added up the concentration of the supply air CO2 concentration.

cexh = csource q̇total

+ csup [4]

After the addition of the supply air CO2 concentration and the exhaust air CO2 concentration to [3], then the effectiveness of personalized ventilation is defined by:



𝜀𝑝,𝑝𝑣 =𝑐𝑒𝑥ℎ,𝑝𝑣−𝑐𝑠𝑢𝑝,𝑝𝑣𝑐𝑝,𝑝𝑣−𝑐𝑠𝑢𝑝,𝑝𝑣

−𝑐𝑒𝑥ℎ,0−𝑐𝑠𝑢𝑝,0𝑐𝑝,0−𝑐𝑠𝑢𝑝,0

𝑐𝑒𝑥ℎ,𝑝𝑣−𝑐𝑠𝑢𝑝,𝑝𝑣𝑐𝑝,𝑝𝑣−𝑐𝑠𝑢𝑝,𝑝𝑣

= 1 −𝑐𝑒𝑥ℎ,0−𝑐𝑠𝑢𝑝,0𝑐𝑝,0−𝑐𝑠𝑢𝑝,0

𝑐𝑒𝑥ℎ,𝑝𝑣−𝑐𝑠𝑢𝑝,𝑝𝑣𝑐𝑝,𝑝𝑣−𝑐𝑠𝑢𝑝,𝑝𝑣

[5]

Where,

cexh,0 = concentration in the exhaust, without PV cp,0 = concentration at point p, without PV csup,0 = concentration of the supply air, without PV cexh,pv = concentration in the exhaust, with PV cp,pv = concentration at point p, with PV csup,pv = concentration of the supply air, with PV

A value of 1 presents a situation where local air quality is comparable to outside air

0 presents a situation where local air quality with PV is comparable to local air quality without PV

>0 presents a situation where local air quality with PV is better than local air quality without PV

<0 presents a situation where local air quality with PV is worse than local air quality without PV

Calculations and post-processing of the results is done in Matlab R2010b (M-files are shown in Appendix J) and Excel 2007.

4. Results

4.1 Static experiments Table 2 shows the local air quality index at the measurement positions. A value of 1 means perfect mixing, a value above 1 means a better local air quality at a specific point in comparison with the exhaust air and a value below 1 means a worse air quality in comparison to the exhaust air. As not all data sets are normally distributed, the median of each data set is taken. Two CO2 measurement series, (Appendix I) have been done twice. Results are all within 10% of each other, so accuracy of the measurement set-up is determined to be 10 %. To reveal the reliability of the measurements a bandwidth of the median ±2σ is calculated, this represents a reliability of 95%. To calculate the bandwidth of the effectiveness of PV with formula [5], the lower range of the situation with PV is compared to the higher range of the situation without PV and vice versa. Table 2: Local air quality [LAQ] index at different positions (Figure 7) for the static experiments. A bandwidth of 95% reliability (𝑀𝑒𝑑. ±2𝜎) is shown in between brackets.

Position 1 LAQ [-]

Position 2 LAQ [-]

Position 3 LAQ [-]

Position 4 LAQ [-]

Position 5 LAQ [-]

Patients lying both face up PV on at both patients 0.58 (0.32; 3.2) 1.35 (1.13; 1.69) 1.35 (1.11; 1.71) 1.53 (1.33; 1.79 1.72 (1.64; 1.81)

PV off 0.94 (0.74; 1.29) 0.96 (0.91; 1.02) 0.97 (0.91; 1.03) 0.97 (0.94; 1.00) 0.98 (0.95; 1.02)

Patients lying face to face PV on at both patients 0.99 (0.82; 1.46) 0.84 (0.71; 1.00) 0.90 (0.81; 0.99) 0.99 (0.84; 1.21) 1.64 (1.54; 1.71)

PV off 0.75 (0.58; 1.07) 0.91 (0.63; 1.64) 0.83 (0.60; 1.37) 0.97 (0.83; 1.16) 0.98 (0.92; 1.04)

Table 3 shows the effectiveness of personal ventilation as described in chapter 3. A value of 0 means no improvement of local air quality due to PV and 1 means outside contamination conditions. Personal ventilation is able to improve the local air quality to nearly outside conditions as the effectiveness is 0.98 (98%) when patients



are lying face up and 0.95 (95 %) when patients are lying face to face. When patients are lying face up, the effectiveness of the PV system is also high for the measurement positions 0.5m above the receiver’s mouth and at the positions in between the beds. When patients are lying sideways, the effectiveness of the PV system is significantly lower or even negative at those positions. Table 3: Effectiveness of personal ventilation system at different positions (Figure 7) for the static experiments. A positive value (max. 1=100% clean air) means an improvement, a negative value means a worsening of the local air quality due to the implementation of PV system for both patients, see Chapter 3. A bandwidth of 95% reliability (𝑀𝑒𝑑. ±2𝜎) is shown in between brackets.

Position Effectiveness of PV for both patients in comparison with no PV

[-]

Effectiveness of PV for both patients in comparison with no PV

[-] Patients lying both face up Patients lying face to face

1. 0.5m above source patient -1.32 (<-2; 1) 0.43 (-0.83; 0.91)

2. Middle, 1.2m above floor 0.63 (0.22; 0.95) -0.17 (<-2; 0.59)

3. Middle, 1.65m above floor 0.62 (0.16; 0.97) 0.16 (<-2; 0.61)

4. 0.5m above receiver patient 0.81 (0.58; 1) 0.05 (-1.21; 0.61)

5. At receiver patient’s mouth 0.98 (0.91; 1) 0.95 (0.83; 1)

For the purpose of estimating the probability of airborne transmission of an infectious person indoors, Riley (Riley et al., 1978) developed the Wells–Riley equation:

𝑃 = 1 − 𝑒−𝑖𝑞𝑝𝑡𝑄 [6]

Where, P = the probability of infection for susceptible population

i = the number of infectors, q = the quantum generation rate by an infected person; exposure to one quantum of infection gives an

average probability of 63% (1-e-1) of becoming infected (essentially an infectious dose of 63%, ID63). p = the breathing rate per person [m3/s] t = the total exposure time [s] Q = the outdoor air supply rate [m3/s]

So the exposure time in the experimental room with only the mixing room ventilation active, where 63% of the humans will be infected is: 𝑡𝑃=0.63 = ln (1 − 0.63) ∙ 𝑄

−𝑖𝑞𝑃=0.63𝑝 [7]

Rudnick (Rudnick & Milton, 2003) used data from the influenza outbreak aboard an aircraft and calculated 79 to 128 quanta per hour for a highly infectious influenza case. Apply this in [7] and the exposure time of a receiver patient to get average infection probability of 63% of the influenza virus in the room without PV can be calculated (Table 4). Multiplying those values with (1-effectiveness_Personal Ventilation) gives the exposure time when PV is active. Table 4: Average exposure time of receiver patient to get an infection probability of 63% for the influenza virus.

Case Exposure time to reach ID63

[hr] PV off, patients lying both face up 4.8 PV off, patients lying face to face 4.8 PV on, patients lying both face up 244 PV on, patients lying face to face 98

4.2 Continuous movements Visualisations of the smoke distribution, spread through the air flow of the PV pillow of the receiver patient are shown in Figure 11, where changing pressure fields at t= 3s and t= 9s and vortices at t= 6s could be observed.



0 s 3 s 6 s 9 s

Figure 11: Visualization of the air flow above the source patient, when a person is moving in between the beds. Start of the moving person is at 0 s, at t=3s, the manikin has reached the position close to the patients’ head, where it stays for 5 s. Table 5 shows the local air quality index at the measurement positions when a manikin is continuously moving in between the beds. Table 6 shows the effectiveness of PV in this situation. Table 5: Local air quality [LAQ] index at different positions (Figure 7) for the continuous movement experiments. A value of 1 means perfect mixing, a value above 1 means a better air quality at a specific point in comparison with the exhaust air and a value below 1 means a worse air quality in comparison to the exhaust air. A bandwidth of 95% reliability (𝑀𝑒𝑑. ±2𝜎) is shown in between brackets.

Position 1 LAQ [-]

Position 2 LAQ [-]

Position 3 LAQ [-]

Position 4 LAQ [-]

Position 5 LAQ [-]

Patients lying both face up PV on at both patients 1.00 (0.42; >3) 1.13 (1.01; 1.26) 1.14 (1.05; 1.25) 1.25 (1.10; 1.43) 1.56 (1.36; 1.83)

PV off 0.91 (0.54; 2.83) 0.94 (0.90; 0.99) 0.95 (0.91; 1.00) 0.96 (0.92; 1.01) 0.96 (0.92; 1.00)

Patients lying face to face PV on at both patients 0.97 (0.79; 1.23) 0.81 (0.72; 0.92) 0.85 (0.79; 0.95) 0.95 (0.79; 1.19) 1.58 (1.47; 1.72)

PV off 0.80 (0.63; 1.07) 0.90 (0.54; 2.78) 0.89 (0.75; 1.10) 0.95 (0.90; 1.01) 0.97 (0.90; 1.05)

Table 6: Effectiveness of personal ventilation system at different positions (Figure 7) for the continuous movement experiments. A positive value (max. 1) means an improvement, a negative value means a worsening of the local air quality due to the implementation of PV system for both patients, see Chapter 3. A bandwidth of 95% reliability (𝑀𝑒𝑑. ±2𝜎) is shown in between brackets.

Position Effectiveness of PV for both patients in comparison with no PV

[-]

Effectiveness of PV for both patients in comparison with no PV

[-] Patients lying both face up Patients lying face to face

1. 0.5m above source patient 0.19 (<-2; 1) 0.33 (-0.97; 0.78)

2. Middle, 1.2m above floor 0.37 (0.05; 0.62) -0.26 (<-2; 0.60)

3. Middle, 1.65m above floor 0.39 (0.12; 0.60) -0.12 (-1.24; 0.37)

4. 0.5m above receiver patient 0.54 (0.22; 0.80) -0.01 (-0.71; 0.52)

5. At receiver patient’s mouth 0.91 (0.66; 1) 0.91 (0.75; 1)

4.2.1 Influence of movements In Table 7 the influence of movements on the local air quality at a specific point around the source and receiver patient (Figure 7) is shown. Results at positions in between the beds (position 2 and 3) are not shown, as the path of the moving manikin crosses out the sensor position. As can be seen in the first column of Table 7 movements do not significantly influence the local air quality when there is no PV system working. When the PV systems are active, large improvements of the LAQ could be seen at the sensor position 0.5m above the source patient. The movements mostly worsen (max. 26 %) the LAQ at 0.5m above the receiver patient, as the worsening at breathing level is lower.



Table 7: Improvement of local air quality due to continuous movement of a walking person in terms of percentage Position No PV

[%]

PV for both patients

[%]

PV only for source patient

[%]

PV only for receiver patient

[%] Patients lying both face up 1. 0.5m above source patient -3.3 72.6 34.7 -16.7

4. 0.5m above receiver patient’s mouth -0.5 -18.4 -13.9 -26.2

5. At receiver patient’s mouth -2.0 -9.4 -17.7 -10.1

Patients lying face to face 1. 0.5m above source patient 5.8 -2.2 26.0 16.1

4. 0.5m above receiver patient’s mouth -1.5 -4.1 6.3 -11.9

5. At receiver patient’s mouth -0.7 -3.5 1.2 -3.7

4.3 Single time movements Moving person: The particle counts at the mouth of the receiver patient, during 10seconds after one movement of a moving person in between the beds is given in Figure 12. Same level peaks can be seen for particles between 0.3 and 0.5 𝜇m as for the particles larger than 0.5 𝜇m. The movement start at t= 0 s and at t= 3 s the moving person has reached the position close to the heads of the patients. Here it pauses for 10 seconds and moves back.

(a) (b) Figure 12: Number of particles at the mouth of the receiver patient, during time after the movement of a person in between the beds, t=0 when person starts moving. a): situation when PV is off (n=56), b): situation when PV is on at both patients (n=50). An average concentration during the whole measurement series is also given. Moving bed-sheets: The particle counts during 10 seconds after a bed sheet is turned over, is given in Figure 14. The movement is started at t=0 s and immediately returns when it has reached its upper position. A visualization of this movement is shown in Figure 13. The pressure wave can be clearly seen, in the first few seconds after the movement, but within 10 seconds the air-flow returns back in original state.



0 s 2 s 3 s

Figure 13: Visualization of the exhalation air flow between the two patients, when bed-sheets are turned over. Patients are lying face to face. The sheet is moved up at t=0 s and returns to original position at t=1.5s. The source patient is positioned at the left and the receiver patient is at the right side of the pictures.

(a) (b)

Figure 14: Number of particles at the mouth of the receiver patient, during time after turning over of a bed-sheet, t=0 when the bed-sheet movement is started. a): situation when PV is off (n=58), b): situation when PV is on at both patients (n=64). An average concentration during the whole measurement series is also given. Opening door: The particle counts during 10seconds after a door is opened, is given in Figure 16. The movement is started at t=0 s and the door is immediately closed. A visualization of this movement is shown in Figure 15. The pressure wave is only small and original state can be seen after 20s.

0 s 3 s 7 s 14 s.

Figure 15: Visualization of the exhalation air flow between the two patients, when door is opened and immediately closed. Patients are lying face to face. Start of the movement is at 0 sec. The source patient is positioned at the left side of the pictures.



(a) (b) Figure 16: Number of particles at the mouth of the receiver patient, during time after opening and immediately closing of the door, t=0 when the door opening is started. a): situation when PV is off (n=54), b): situation when PV is on at both patients (n=55). An average concentration during the whole measurement series is also given.

5. Discussion

5.1 Method This study focuses on the contaminant distribution from breathing zone to breathing zone. Respiratory tract infection is reported as the most common infection way in Dutch hospitals (van der Kooi et al., 2010), however this way of transport is only one substantial way of transport of infectious aerosols, where skin infections, wound infections and surface infections around the patient are also important ways of contaminant distribution through the air. Besides that, infectious aerosols have to penetrate the respiratory tract for a certain distance to cause infection (Morawska et al., 2009). For example a virus causing influenza landing on skin would not cause an infection, landing on the mucous membrane in the nose, it might cause infection. Once on a susceptible site, invading micro-organisms have to overcome a variety of host resistance mechanisms; the higher the number of potential invaders, the more likely they are to defeat host resistance (CIBSE et al., 2005). Although larger droplets have higher contaminant-carrying capacities, their ability to penetrate deep into the respiratory tract is lower than that of smaller droplets. Smaller pathogen-laden droplets have higher infectivity compared to the larger ones for diseases that have the lower respiratory tract as the target infection site (Nicas et al., 2005). Mouth breathing was used during all tests, another study by Qian (Qian et al., 2006) reveals that the personal exposure level of the receiving manikin will be lower when nose breathing for the source manikin is used. During this research there has been made use of different measurement techniques. The static and continuous experiments are done with CO2 tracer gas. CO2 has got comparable density parameters as the surrounding air and therefore moves comparable. Where the available CO2 sensors did not have a fast enough response time, the single time movements are done with smoke gas. The detected particles are larger than 0.3 𝜇m and thus behave different from CO2 gas in relation to gravitational and momentum forces. The sizes of aerosolized viruses bacterial and fungal spores are 0.02 to 0.30 𝜇𝑚; 0.3 to 10 𝜇𝑚 and 2.0 to 5.0 𝜇𝑚 respectively (Cole & Cook, 1998). However several studies (Loomans & Lemaire, 2002; Nielsen, 2009; Chen & Zhao, 2010) validate the use of smaller particles as tracer gas to simulate the dispersion of small droplet nuclei (max 10 𝜇𝑚). However these validations of different particle sizes are dependent on room ventilation rates. A considerable amount of data sets is not normally distributed (Appendix I). One reason for this are the peaks of CO2 concentration which occur in an irregular base especially in the extension of the exhalation jet. As can be



seen in Figure 17 the exhalation jet of the source manikin can easily move a few degrees due to the smallest disturbance in the air velocity field and this displacement could mean that the measurement point is just inside or just outside the exhalation jet with high concentration of the tracer gas. Concentrations of the tracer gas are also measured with an independent sensor nearby measurement point 1 and similar variations are seen (Appendix I). As not all data sets are normally distributed, the median is used to calculate the local air quality and the effectiveness of the PV. The bandwidth with a reliability of 95% at the positions in the range of the exhalation jet is large, so more measurements are needed to get more reliable results. The bandwidth in the breathing zone of the receiver patient is smaller and so reliable conclusions could be made at this position.

Figure 17: Varying vector field of the exhalation plume of the source manikin under steady state conditions. The picture on the right is taken two seconds after the left picture. With the source patient lying in bed facing sideways, the body thermal plume is weak and so the exhalation does not necessarily follow the boundary layer close to the body, but is able to break free and penetrate the breathing zone of other persons (Qian et al., 2006). During this research there has not been made use of a breathing mechanisms which inhales and exhales several times per minute. This is done to simplify the model as in this case, the breathing process is not interfering with the movements and continuously exhalation may be seen as the worst case scenario in infection spread. However the breath process has little effect on the airflow, because the momentum of exhaled air by breathing is relatively small (Chen & Zhao, 2010) and so these measurements are a good representation of the reality. All experiments have been done with a total air change rate of 4 air changes per hour. This air exchange rate is higher than the minimum compulsory ventilation rate of 1.3 ACH according to the Dutch building code, Bouwbesluit (Bouwbesluit, 2003). However the international accepted building standards , CIBSE (CIBSE et al., 2005), ASHRAE (ASHRAE, 2003) and the American institute of architects advice a higher ventilation rate of 6 ACH. The used ventilation rate of 4 ACH is in between those recommendations and is used in several other comparable patient room ventilation studies (Bjorn and Nielsen, 2002; Qian et al., 2006; Qian et al., 2008; Mazumdar et al., 2010). It is generally accepted that higher ventilation rates decrease the infection risk (Qian & Li, 2010). In Dutch hospitals there is no uniform ventilation rate, based on performance demands and energy strategies the ventilation rates vary from the minimum compulsory ventilation rate of 1.3 ACH to a multiple of that. The total air change rates in the room remains constant at 4 ACH. Personal ventilation with lower room ventilation rates may also perform well on cross-contamination and this seems interesting for the building industry as a ventilation configuration with lower and better directed airflow rates is less energy consuming and therefore decreases the exploitation costs. A reduction of 1 m3 ventilation air per hour will save 25 € per year (Trip & Bielleman, 2012), so a reduction in ventilation rate of 1 ACH could easily lead to a saving of more than a hundred thousand Euro on yearly base in a 100-room ward department. However this is not tested in this study and effects of lower room ventilation rate should be investigated in further research. For example the conditions for visitors and healthcare workers in the patient room could decreases with a lower room ventilation rate.

Sensor Sensor

Mouth exhalation

Mouth exhalation

Exhalation cloud

Exhalation cloud



Thermal comfort The thermal comfort of the measured PV system has not been specifically investigated during this study. Other studies of personal ventilation (Cermak et al., 2006) reveal that personal ventilation is able to improve thermal comfort as a personal micro-atmosphere can be created around the patient. The breathable pillows used in this study are not optimized for creating thermal and lie comfort, but are only designed to verify the performance on infection control. However some experiments are done to quantify the air velocity and noise production of the pillows. The outlet air velocity through the fabric could be calculated by:

𝑣 = �̇�𝐴

= 0.0150.425

= 0.04 𝑚𝑠

[8]

Where, v= Air velocity in m/s �̇�= Air flow through the pillow in m3/s A= Area of the top of the pillow in m2,

An assumption hereby is made that the area of the air flow openings is the same as the area of the whole breathable top of the pillow. In practice the area of the non permeable part of the pillow fabric could not be neglected and air velocity measurements just above the pillow show air velocities of 0.12 m/s. However this is still under the accepted comfort levels of 0.2 m/s (CIBSE, 2006). However, comfort levels of building services systems are meant on room level and discomfort can still occur because the head of the patient is situated close to the PV system. The noise level rises from 37 dB when the PV pillows were off to maximum 43 dB when the pillows were active. However at heights from 5 cm of the pillow the raising is maximum 2 dB. See Figure 18. According to the Dutch building legislation (Bouwbesluit, 2003) an installation may only produce a sound level of 30 dB. Accumulating with the surrounding sound, the total sound level may not exceed 38 dB. This means that the PV pillow in this configuration is not permitted, however the tested pillow is not optimized for noise reduction and improvements could be made on for example the supply valves.

(a) (b) Figure 18: (a): Velocity field at a specific height z above the width of the pillow. Measured velocities when the pillow was set of are; 0.02 m/s at z = 1cm and 0.04 m/s at other heights. (b): Noise level field at a specific height z above the width of the pillow. Measured noise level when the pillow was off is 37 dB at all points.

5.2 Static situation At breathing level of the receiver patient the effectiveness of the personal ventilation system is up to 98 %, this means that PV is able to reduce contaminant concentration to only 2 % in comparison to a comparable situation where no PV is used. This is in line with results of a PV pillow tested in a displacement ventilated room (Figure 4 )(Nielsen et al., 2007b). PV performance at breathing level of the receiver patient when PV is active at both patients or only at receiver patient is comparable of patients lying face to face and face up.



Significant differences occur on other measurement-points when patients are lying both face up or both sideways face to face. The protective effect of the personal ventilation is still present 0.5m above the receiver patient when both patients are lying face up, as the effectiveness is still 81% there. However, when patients are lying sideways the effectiveness significantly decreases to nearly 0 % at this position. The drop at this position appears to be mainly contributed by the PV pillow at the source patient as results of the case where only the source patient has an PV pillow shows an effectiveness of -47% and the case where only the receiver patient has an PV pillow shows an effectiveness of +68% (see Table 8, Table 9, Appendix G). At health care workers or visitors position, in between the beds, a same trend can be seen. Effectiveness results of 61% when patients lie face up to negative (-17%) when patients are lying face to face. Table 8: Effectiveness of personal ventilation system at different positions (Figure 7), for the static experiments. A positive value (max. 1=100% clean air) means an improvement, a negative value means a worsening of the local air quality due to the implementation of PV system for the receiver patient only, see Chapter 3.

Position

Effectiveness of PV for receiver patients only in comparison with no PV [-]

Face up Face to face 1. 0.5m above source patient 0.19 0.05 2. Middle, 1.2m above floor 0.38 -0.94 3. Middle, 1.65m above floor 0.35 -0.08 4. 0.5m above receiver patient 0.75 0.68 5. At receiver patient’s mouth 0.98 0.98

Nielsen (Nielsen et al., 2007b) reveals with CFD simulations that a PV pillow may be used to remove exhalation from a source when it is applied in connection with uni-directional ventilation. Because of the implementation of the PV system and helped by the buoyancy effect, the contaminants could rise to the ceiling of the room and could be removed there. This is especially important for healthcare workers and visitors in the patient room, because the PV system does not protect them. The concentration level at the neighbouring bed close to the source manikin is reduced to a factor of 0.6 when only the source manikin uses a pillow (Nielsen, 2009). Table 9 shows that this effect is observable with mixing ventilation, when patients are lying both faces up as effectiveness of the source patient pillow is around 0.4, which means a contaminant concentration reduction of 40% on healthcare worker and receiver patient position. However when patients are lying sideways, the effectiveness reduces dramatically to negative values, what means that the contaminant concentration is worse in comparison to a situation where no PV is used.

Table 9: Effectiveness of personal ventilation system at different positions (Figure 7), for the static experiments. A positive value (max. 1=100% clean

Figure 19: The distribution of contaminants from a source manikin. The room is ventilated by displacement ventilation, and the source manikin is ventilated by personal ventilation pillow (Nielsen et al., 2007b).

air) means an improvement, a negative value means a worsening of the local air quality due to the implementation of PV system for the source patient only, see Chapter 3. The room is ventilated with mixing ventilation.

Position

Effectiveness of PV for source patient only in comparison with no PV [-]

Face up Face to face 1. 0.5m above source patient -1.09 0.04 2. Middle, 1.2m above floor 0.36 -0.97 3. Middle, 1.65m above floor 0.37 -0.37 4. 0.5m above receiver patient 0.41 -0.47 5. At receiver patient’s mouth 0.45 -0.37



The average exposure time of receiver patient to get an infection probability of 63% for the influenza virus increases multiple times from 7.3 hours when no PV is used to 146-365 hours when the PV system is active (Table 4). This means that the average risk to get infected through a nearby infectious patient is minimized by use of the PV pillows. However patients must be bed-bound to get these high results and patients at risk like elderly people and people with a low immune system get infected earlier. This method does not consider; the decay of viability of the virus, the size of the aerosols and the dispersion due to coughing or sneezing, so results may be overestimated. Health care workers are less protected by a personal ventilation pillow, as infection probability times increase maximum 39% and could even be negative in some situations. This study gives a risk analysis by comparing contaminant concentration and local air quality in several situations, however in infection control it is important to take in mind that in some cases only one single infectious particle can cause an infection.

5.3 Continuous Movements When a person is moving along a bed side, vortices and pressure areas are generated that may transport air and contaminant from one patient to another (Figure 11). When a person is moving it will push the air with its contamination away perpendicular to the direction of moving. Contaminants floating in between the beds are forced into the direction of the beds and the momentum that these contaminants receive due to this action may not be interrupted by the personal ventilation system and penetrate the breathing area. Vortices increase the unpredictability of the air flow and cause a hard controllable system. Concentration peaks are quick removed from the breathing zone of the receiver patient, as the breathing zone is cleared by the PV quicker as the surrounding air further away from the pillow (Figure 20).

0 s 3 s 9 s

Figure 20: Contaminant reduction of the breathing zone after an intervention of a contaminant cloud. White arrows visualize the flow direction of the contaminants, blue arrows visualize clean air from the PV pillow. The influence of a continuous moving person is substantial, as it increases contaminant concentration at breathing level or 0.5m above breathing level of the receiver patient, when patients are lying both face up and personal ventilation is used (Table 7). When there is no personal ventilation installed there is no significant difference visible. Comparison of the local air quality at 0.5 m above breathing level reveals that the contaminant concentration is actually lower with movements when the personal ventilation is switched on at the same patient. The reason for this is that the movement disperses the uni-directional contaminant transport from the pillow and this may be an advantage for patients who are in potential risk by exposure to their one exhalation air like anaesthetic patients recovering from a surgery, whereas a floating away of narcotic gasses is connected to a quick recovery. During this research the moving manikin is simulated by a manikin without arms and legs, on strict path on a rail parallel on the bed side. Results could be influenced by the quick movement of limbs nearby a patient or due to movements of a person perpendicular to the bed direction. Similarly the shape of the used manikin is not exactly the shape of a representative walking person and the height of 1.9m is not representative for the average population and may overestimate the results. It is also frequently occurring in a patient room situation that more than one person is moving around the bed, for example the cooperation of multiple nurses in a nursing operation



or the visiting of several visitors. This will increase the simultaneous movements in the room and it is therefore plausible that this will substantially increase the air movements and therefore the contaminant distribution.

5.4 Single time movements For every experiment two independent test series are done on different shifts during a day and afterwards results of both series are compared to get the reliability of the tests. All tests are within 20% of each other, so the reliability of these experiments is determined to be 20%. In Figure 12 can be seen that in the situation when the PV is on and a moving person is walking in between the beds, an increase of particles of more than 5 times the average concentration is counted in the period between 10 and 20 seconds after the start of the movement. There could also be observed a substantial increasing of particles between 20 and 30 seconds, while thereafter no substantial increasing could be seen. This is substantially more than Mazumdar et al (Mazumdar et al., 2010) reported as he concluded that the variation for all movements was within 25%. However Mazumdar used CFD simulations with other boundary conditions, so results are not directly comparable. The increasing due to the movement in this study looks dramatic, but the concentration is taking into account of the error limits still lower compared to the average concentration measured in the same sort of tests, when the PV is off. In this last situation only a relatively small increase of particle counts between 10 and 20s after movement can be seen. During many moments in patient rooms, there is not only one person moving in between the beds, but several healthcare workers or visitors could walk around, this study did not consider the interfering effect of multiple movements at the same time. No substantial difference in particle counts can be seen after the turning over of a bed sheet (Figure 14). In the visualization (Figure 13) can be seen that there is an obvious air movement due to the movement, but that this air movement reaches maximum 1m, and so the particle differences after one meter, where the receiver patient breaths, is minimal. The turning over of a bed sheet tested here is a symmetrical movement, similar to when bed-sheets are turned over with two hands at the whole width of the bed-sheet. Turning over of bed-sheets at one side, may causes a more directed but also less powerful airflow. Door opening does not cause a significant difference in particle counts at the receiver patient (Figure 16). In the visualization (Figure 15) can be seen that the momentum of the moving air is visible nearby the source patient, but that this momentum is too low to reach the breathing zone of the receiver patient. By opening the test door clean air with other temperature characteristics may enters the room, this may influences the results. The test door opened to the outside, in hospitals most doors open to the inside of the room are sliding doors are used; results of other door types may differ.

6. Conclusion To answer the research question how personal ventilation systems perform on infection control in comparison to non personal systems in hospital patient rooms, it can be concluded that personal ventilation systems perform better in contaminant reduction for the receiver patient’s breathing zone in both static, continuous dynamical situations as well as single time movements. From the results of the full-scale static experiments it can be concluded that the personal ventilation pillow works effective on contaminant reduction on the receiver patient in a mixed ventilated room as an effectiveness of the PV pillow of 98% (83% -100%; p=0.05) is achieved. The PV pillow performs similar in protection of the receiver patient against airborne cross infection at mouth level, when patients are lying on their back and when patient are lying sideways face to face. A PV pillow at the source patient is not able to reduce emission from the source patient in a mixed ventilated room. When patients are lying sideways and PV is only active at the source patient, a decrease of local air quality is observed at healthcare workers position and at the receiver patient. So, to prevent for infection spread



in a multi person patient room, it is not recommended to provide a PV pillow to a (possible) infectious patient only. A continuous movement of a walking person decreases the local air quality with maximum 26 % at breathing level or at 0.5m above breathing level of the receiver patient, when both patients are lying face up and personal ventilation is used. The PV is only partly capable to bend away the momentum of particles, which is created by the vortices and pressure fields caused by the movement. However during continuous movements PV is still able to reduce contaminant concentration with 91% (66% - 100%; p=0.05) at the breathing zone of the receiver patient. The increase of particles due to a single time moving person is substantial higher (5 times) when PV is on, where only a small increase can be seen when PV is off. However peak concentrations when PV is on are still a factor 2 lower than the average concentration when PV is off. Particle concentrations return to normal within 30 s after the start of the moving manikin. No significant variations on particle counts at the breathing zone of the receiver patient are observed when bed sheets are turned over or when a door is opened and closed. Movements do only cause significant differences in contaminant concentration when the movement is nearby the investigated zone.

7. Further research More research is needed on the subject of comfort of patients lying on a breathable pillow, since a breathable pillow creates a light breeze around the patients head and a light intensified sound level. Before operation of the breathable pillows, prolonged test should be done about the comfort of patients in both awake and sleeping situations. This includes the application of different supply air temperatures. For economical and energy saving reasons, the personal ventilation system should be tested with several room air supply rates. Personal ventilation is maybe able to decrease the total ventilation rate in a hospital patient room without losing its function on infection control, but further research is needed on this subject. Especially a risk assessment on visitors and health care workers should be done in that case. Research with variable supply air flow rates through the pillows is also recommended. Smart systems could anticipate on movements and adapt the ventilation system to minimize the infection risk.

8. Acknowledgements This research is done in cooperation with BAM techniek and would not have been possible without the facilities at the TU/e building physics and services laboratory.



References AIA (2006). Guidelines for design and construction of health-care facilities/ publ. of the American institute of

Architects. New York 2006: American institute of Architects.

ASHRAE (2003). ASHRAE Handbook, HVAC applications, SI edition; Atlanta : ASHRAE, 2003

Beggs, C. B. (2003). The airborne transmission of infection in hospital buildings: Fact or fiction? Indoor and Built Environment, 12, 9-18.

Bjorn, E. & Nielsen, P. V. (2002). Dispersal of exhaled air and personal exposure in displacement ventilated rooms. Indoor Air, 12, 147-164.

Borkow, G. & Gabbay, J. (2008). Biocidal textiles can help fight nosocomial infections. Medical Hypotheses, 70, 990-994.

Bouwbesluit. (2003). Bouwbesluit 2003 presented by Bris Warenhuis Rotterdam. http://www.briswarenhuis.nl/

Cermak, R., Melikov, A. K., Forejt, L., & Kovar, O. (2006). Performance of personalized ventilation in conjunction with mixing and displacement ventilation. Hvac&R Research, 12, 295-311.

Chen, C. & Zhao, B. (2010). Some questions on dispersion of human exhaled droplets in ventilation room: answers from numerical investigation. Indoor Air, 20, 95-111.

CIBSE, A. (2006). CIBSE guide A, Environmental Design. (7th edition ed.) London: CIBSE.

CIBSE, B., Barnard, N., & Jaunzens, D. (2005). CIBSE guide B: Section 2: Ventilation and air conditioning. In ( London: CIBSE.

Cole, E. C. & Cook, C. E. (1998). Characterization of infectious aerosols in health care facilities: An aid to effective engineering controls and preventive strategies. American Journal of Infection Control, 26, 453-464.

Eames, I., Shoaib, D., Klettner, C. A., & Taban, V. (2009). Movement of airborne contaminants in a hospital isolation room. Journal of the Royal Society Interface, 6, S757-S766.

Hathway, E. A., Noakes, C. J., Sleigh, P. A., & Fletcher, L. A. (2011). CFD simulation of airborne pathogen transport due to human activities. Building and Environment, 46, 2500-2511.

Hopmans, T. E. M., Blok, H. E. M., Troelstra, A., & Bonten, M. J. M. (2007). Prevalence of hospital-acquired infections during successive surveillance surveys conducted at a university hospital in the Netherlands. Infection Control and Hospital Epidemiology, 28, 459-465.

Klettner, C. A., Eames, I., & Tang, J. W. (2009). The effect of turbelence on the spreading of infectious airborne droplets in hospitals. In Synthetic models of turbelence (pp. 141-152). Berlin, Germany: Springer Verlag.

Knibbs, L. D., Morawska, L., Bell, S. C., & Grzybowski, P. (2011). Room ventilation and the risk of airborne infection transmission in 3 health care settings within a large teaching hospital. Am.J.Infect.Control.

Kowalski, W. J. (2002). Immune building systems technology. New York: Mc Graw Hill.

Li, Y., Chan, A., Leung, K., & Lee, J. H. W. (2003). Dispersion and control of SARS virus aerosols in indoor environment - Transmission routes and ward ventilation Hong Kong: University of Hongkong.

http://www.briswarenhuis.nl/�



Li, Y., Leung, G. M., Tang, J. W., Yang, X., Chao, C. Y. H., Lin, J. Z. et al. (2007). Role of ventilation in airborne transmission of infectious agents in the built environment - a multidisciplinary systematic review. Indoor Air, 17, 2-18.

Loomans, M. G. L. C. & Lemaire, T. (2002). Particle concentration calculations using CFD. In Proceedings Roomvent 2002, Copenhagen Denmark, 393-396.

Mazumdar, S., Yin, Y. G., Guity, A., Marmion, P., Gulick, B., & Chen, Q. Y. (2010). Impact of Moving Objects on Contaminant Concentration Distributions in an Inpatient Ward with Displacement Ventilation. Hvac&R Research, 16, 545-563.

Melikov, A. K. (2004). Personalized ventilation. Indoor Air, 14, 157-167.

Melikov, A. K., Cermak, R., & Majer, M. (2002). Personalized ventilation: evaluation of different air terminal devices. Energy and Buildings, 34, 829-836.

Morawska, L., Johnson, G. R., Ristovski, Z. D., Hargreaves, M., Mengersen, K., Corbett, S. et al. (2009). Size distribution and sites of origin of droplets expelled from the human respiratory tract during expiratory activities. Journal of Aerosol Science, 40, 256-269.

Nicas, M., Nazaroff, W. W., & Hubbard, A. (2005). Toward understanding the risk of secondary airborne infection: Emission of respirable pathogens. Journal of Occupational and Environmental Hygiene, 2, 143-154.

Nielsen, P. V. (2009). Control of airborne infectious diseases in ventilated spaces. Journal of the Royal Society Interface, 6, S747-S755.

Nielsen, P. V., Hyldgaard, C. E., Melikov, A., Andersen, H., & Soennichsen, M. (2007a). Personal exposure between people in a room ventilated by textile terminals - with and without personalized ventilation. Hvac&R Research, 13, 635-643.

Nielsen, P. V., Jiang, H., & Polak, M. (2007b). Bed with Integrated Personalized Ventilation for Minimizing Cross Infection. In Proceedings Room Vent 2007 Helsinki, Finland.

Nielsen, P. V., Polak, M., Jiang, H., Li, Y., & Qian, H. (2008). Protection against cross infection in hospital beds with integrated personalized ventilation. In Indoor Air 2008 Copenhagen: Denmark.

Pantelic, J., Sze-To, G., Tham, K. W., Chao, C. Y. H., & Khoo, Y. C. M. (2009). Personalized ventilation as a control measure for airborne transmissible disease spread. Journal of the Royal Society Interface, 6, S715-S726.

Qian, H. & Li, Y. (2010). Removal of exhaled particles by ventilation and deposition in a multibed airborne infection isolation room. Indoor Air, 20, 284-297.

Qian, H., Li, Y., Nielsen, P. V., & Hyldgaard, C. E. (2008). Dispersion of exhalation pollutants in a two-bed hospital ward with a downward ventilation system. Building and Environment, 43, 344-354.

Qian, H., Li, Y., Nielsen, P. V., Hyldgaard, C. E., Wong, T. W., & Chwang, A. T. Y. (2006). Dispersion of exhaled droplet nuclei in a two-bed hospital ward with three different ventilation systems. Indoor Air, 16, 111-128.

Riley, E. C., Murphy, G., & Riley, R. L. (1978). Airborne Spread of Measles in A Suburban Elementary-School. American Journal of Epidemiology, 107, 421-432.

Rudnick, S. N. & Milton, D. K. (2003). Risk of indoor airborne infection transmission estimated from carbon dioxide concentration. Indoor Air, 13, 237-245.



Shih, Y. C., Chiu, C. C., & Wang, O. (2007). Dynamic airflow simulation within an isolation room. Building and Environment, 42, 3194-3209.

Shiomori, T., Miyamoto, H., & Makishima, K. (2001). Significance of airborne transmission of methicillin-resistant Staphylococcus aureus in an otolaryngology-head and neck surgery unit. Archives of Otolaryngology-Head & Neck Surgery, 127, 644-648.

Tang, J. W., Li, Y., Eames, I., Chan, P. K. S., & Ridgway, G. L. (2006). Factors involved in the aerosol transmission of infection and control of ventilation in healthcare premises. Journal of Hospital Infection, 64, 100-114.

Tang, J. W., Noakes, C. J., Nielsen, P. V., Eames, I., Nicolle, A., Li, Y. et al. (2011). Observing and quantifying airflows in the infection control of aerosol- and airborne-transmitted diseases: an overview of approaches. Journal of Hospital Infection, 77, 213-222.

Trip, A. & Bielleman, R. (8-6-2012). Conversation about hospital ventilation in Meander Medisch Centrum Amersfoort. 8-6-2012. Ref Type: Personal Communication

van der Kooi, T. I. I., Mannien, J., Wille, J. C., & van Benthem, B. H. B. (2010). Prevalence of nosocomial infections in The Netherlands, 2007-2008: results of the first four national studies. Journal of Hospital Infection, 75, 168-172.

Wang, J. L. & Chow, T. T. (2011). Numerical investigation of influence of human walking on dispersion and deposition of expiratory droplets in airborne infection isolation room. Building and Environment, 46, 1993-2002.

Yam, R., Yuen, P. L., Yung, R., & Choy, T. (2011). Rethinking hospital general ward ventilation design using computational fluid dynamics. Journal of Hospital Infection, 77, 31-36.

Zhao, B., Zhang, Y., Li, X. T., Yang, X. D., & Huang, D. T. (2004). Comparison of indoor aerosol particle concentration and deposition in different ventilated rooms by numerical method. Building and Environment, 39, 1-8.



Appendix A: Introduction, Nosocomial infections Nosocomial infections transmitted by air Most nosocomial infections (NI) have been identified as having at least some potential for airborne transmission although most of them are primarily spread by other routes, such as contact (Kowalski, 2009). It has been estimated that the airborne route of transmission accounts for between 10 and 20% of endemic nosocomial infections (Beggs, 2003) (crossreference Brachman 1974). Most airborne micro-organisms are generated within the building by the staff, patients and visitors. Only a minority of the micro-organisms in the air, usually fungal spores originate outside. During human expiratory activities such as talking, laughing, coughing and sneezing many droplets are expelled from the respiratory tract. It is known that respiratory infections can be spread by these droplets and their residues after evaporation (Xie et al., 2009). However also dust particles and skin squamae carrying pathogenic micro-organisms could spread infection by the airborne route and fungal spores are also widely disseminated via the airborne route. When droplets reach a surface, they could still infect people by the contact route, or resuspend by little air movements, gravitational, thermal, molecular or electrical movements. Note also that although the viral particles are themselves small, the size of airborne particles could be greater, as the droplet could exist of other materials. Virtually all infectious agents that can cause infection at long physical distance range can also cause infection at short range as well as direct contact (Tang et al., 2006). Table A.1: Potential airborne nosocomial pathogens. References:(Schaal, 1991) (Beggs, 2003; Tang et al., 2006) Viruses Disease/ Symptoms Aerosol route of transmission Adenoviruses Respiratory tract infections Transmitted through respiratory droplets Coronaviruses Common cold Presumably inhalation of airborne droplets (Coxsackieviruses) Aerosol droplet spread Influenza viruses Flu Airborne spread predominates Measles Airborne by droplet spread Norwalk Virus Airborne transmission from vomits Parainfluenza viruses Croup, pneumonia Airborne spread Parvoviruses 5th disease Contact of infected secretions Respiratory syncytial virus Cold in the nose Presumably inhalation of airborne droplets Rhinoviruses Common cold Presumably inhalation of airborne droplets Rotavirus Vomiting/ diarrhoea Possible respiratory spread, aerosols from faeces and vomits Rubella Droplet spread Varicella-zoster virus Chicken pox Droplet spread of vesicle fluid or respiratory tract secretions Bacteria

Acinetobacter spp. Bacillus anthracis Anthrax Inhalation of spores Bordetella pertussis Whooping cough Direct contact with discharges from respiratory mucous membranes of

infected persons by the airborne route Corynebacterium diphtheria Diphtheria Enterobacteriaceae Aerosol droplet spread Legionellae Legionellosis Epidemiological evidence supports airborne transmission (Haemophilus infuenzae) Meningitis Droplet infection and discharges from nose and throat (Klebsiella pneumoniae) Klebsilla (Mycoplasma pneumoniae) Pneunomia Probably droplet inhalation Mycobacterium tuberculosis Tuberculosis Exposure to tubercle bacilli in airborne droplet nuclei Neisseria meningitides Meningitis Respiratory droplets from nose and throat Nocardia spp. Acquired through inhalation Pseudomonas spp. Droplet spread Staphylococcus aureus Airborne spread (Streptococcus pneumoniae) Streptococcus Droplet spread and contact with respiratory secretions Streptococcus pyogenes Streptococcus Large respiratoty droplets Fungi Aspergillus spp. Aspergillus Inhalation of airborne conidia (Mucoraceae)

http://nl.wikipedia.org/wiki/Bacillus_anthracis�



Prevalence of nosocomial infections In Dutch hospitals, surveillance for nosocomial infections became mandatory in 1994 (RIVM, 2008). A list is made of 42 diseases which are; highly infectious, have international consequences, or reports of diseases which are necessarily needed for medical treatment or prevention. Not all of these 42 infection diseases use the air as transmission route. In Table A.2 results of the reports of diseases, which could be airborne are summarized for the period 2001-2010. Note that the reports take note of infection diseases in all traits. They could be acquired everywhere and may not have to be specifically acquired in a hospital. Information on what the exact location and time of the infection source is, is medically hard to get and does not exist for most infection diseases. Table A.2: Reports of infection diseases in hospitals. Spreading by air is substantial occurring in all described infection diseases, but doesn’t have to be the main route of transmission. References:RIVM; Melden van infectieziekten, conform de Wet publieke gezondheid 2008 (RIVM, 2008)and RIVM Bulletins infectieziekten (RIVM, 2011)

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Average (2001-2010)

SARS 0 0 0 0 0 0 0 0 0 0 0 Diphtheria 0 0 0 0 0 0 0 0 0 0 0 Pertussis/ whooping cough

6986 5877 2652 89$28 6759 4164 7374 8707 6503 4303 6225

Legionellosis 182 288 195 241 280 455 322 341 240 456 300 Measles 17 3 3 12 3 1 4 109 11 19 18 Rubella 4 3 1 34 362 13 4 2 7 0 43 Meningococcus 770 655 356 305 260 168 195 162 153 144 317 Tuberculosis 1492 1415 1315 1324 1155 1021 960 999 951 941 1157 Anthrax 0 0 0 0 0 0 0 0 0 0 0 New Influenza A (H1N1)*

- - - - - - - - 3413 541 1977

Aviar Influenza* - - - - - - - - 0 0 0 Influenza B* - - - - - - - - 0 45 23 Streptococcus* - - - - - - - - 252 217 235 MRSA* - - - - - - - 2693 2969 3130 2931 *) Duty to report from 1-12-2008, according to "Wet publieke gezondheid". Since 1994, many Dutch hospitals have performed prevalence studies, but only a few have published their data. Many studies focus on the impact of nosocomial infections in the ICU, however a Belgian study in 2010 (Gordts et al., 2010) showed that 81% of the NI’s were observed in non-ICU patient rooms. Taken into account that patients are frequently transferred between patient rooms; prevalences are therefore mainly determined at department level instead of the patient room level (Hopmans et al., 2007). Hopmans studied 2 Dutch hospitals during 2001-2004 and observed 340 Nosocomial infections in 295 (11.1%) of 2661 patients surveyed. Respiratory tract infection (RTI) was the most common infection [22.6%], followed by surgical site infection (SSI) [19.4%]. A prevalence study under 26937 patients in Dutch hospitals, showed a nosocomial infection rate of 6.1% (van der Kooi et al., 2010), where a study done in the mid nineties showed a NI rate of 14 % (Kamp-Hopmans et al., 2003), see table A.4. There is no data available for the role airborne transmission plays in nosocomial infections in the Netherlands. Other European prevalence surveys give comparable results. Average prevalence of NI per 100 admissions is 7.0 in Spain in 1997 (Vaque et al., 1999), 9.1 in Greece in 1999 (Starakis et al., 2002), 8.0 in Denmark in 1999 (Christensen & Jepsen, 2001), 5.1 in Norway in 2002(Eriksen et al., 2005) and 7.1 per 100 in Belgium in 2007(Gordts et al., 2010). Infectious dose To determine the dose or number of microorganisms that will cause infections in 50 percent of an exposed population, the mean infectious dose ID50 is introduced. Units for ID50 are always in terms of microorganisms, or more correctly in the number of colony forming units per cubic meter (cfu/m3). Colony forming units are the number of colonies of a microorganism that grow on plates or petri dishes after sampling a volume of air (Kowalski, 2002). In the same way the mean lethal dose LD50 is constructed.



Elderly, ill people or people with a weak immune system will acquire infections at considerable lower rates than the normal curve of the ID50 (Kowalski, 2002). So hospitalized patients are more susceptible to get infected than the listed ID50 values, however ID50 data for people with a weak immune system are not available. The dose from exposure to airborne pathogens applies to the inhaled does. However many microbial pathogens without an infection resulting are encountered. They have to reach a susceptible site; a virus causing influenza landing on skin would not cause an infection, landing on the mucous membrane in the nose, it might cause infection. Once on a susceptible site, invading micro-organisms have to overcome a variety of host resistance mechanisms; the higher the number of potential invaders, the more likely they are to defeat host resistance. (CIBSE et al., 2004). The infectious dose for Norwalk Like Viruses is as low as 10-100 particles, while approximately dozens of millions particles are spread by one vomit bolus (Caul, 1994). Table A.3: Infectious doses

Pathogen ID50 [CFU/m3]

LD50 [CFU/m3]

Source

Bacillus Antracis 10000 28000 (Kowalski, 2002) Bordetella pertussis 4 1314 (Kowalski, 2002) Coxiella burnetti (Q fever) 1-10 - (Franz et al., 1997) Ebola and Marburg virus 1-10 - (Franz et al., 1997) Influenza A 20 - (Kowalski, 2002) Legionella Pneumophila <129 140000 (Kowalski, 2002) Mycobacterium Tuberculosis 1-10 - (Kowalski, 2002) Norwalk like virus 10-100 - (Caul, 1994) Smallpox 10-100 - (Franz et al., 1997)

Prolonged length of stay Prolongation of stay is one of the most important variables used to measure costs related to nosocomial infections. It is measurable and related to all other cost factors (Erbaydar et al., 1995). A study on surgical wards in a Dutch university hospital (Kamp-Hopmans et al., 2003) showed a significantly longer length of stay for patients with a nosocomial infection than for patients without. Patients with nosocomial infections were hospitalized for 19.8 days versus 7.7 days for patients without nosocomial infections. So patients who developed nosocomial infections (14%) accounted for 30 % of all bed days, increasing the hospitalization costs and the pressure on waiting lists. Hopmans (Hopmans et al., 2007) concluded that patients with a NI had a longer length of stay (29.8 days vs. 17.5 days; P<0.001). The additional length of stay days are divided in ICU and ward days, where the partition is 1.5% versus 98.5% respectively for these locations. However patients with nosocomial infections are often more severely ill and thus are destined to have a prolonged length of stay already, therefore the prolonged length of stay for nosocomial infections is overestimated (Haley, 1991). To reduce the effects of potentially confounding factors such as underlying diseases or age, examinations are done were infected patients are matched with patients who did not acquire a nosocomial infection during their stay. During a two year period, 1482 inpatients were observed in a Turkish hospital (Erbaydar et al., 1995), the results of this examination were an average prolongation of stay per patient of 17 days for 225 infected patients, whereas after matching this number decreased to 11 days. This difference in the results obtained for the matched and unmatched groups supports the idea that patients with severe underlying disease usually had an increased length of stay and thus increase the mean of the group. Economical costs of NI It is important for hospital boards and insurance companies that cost estimates are made for the savings which infection control programs could achieve and therefore scientific evidence should be given to the economical costs of NIs. Studies on costs should estimate both extra length of stay and extra costs attributable to infectious complications (Haley, 1991). Superficial surgical site infections in the Netherlands have been associated with additional costs of € 900 to € 2700 per infection and deep surgical site infections as high as € 1990 to € 3200 (van der Kooi et al., 2010). Data on the extra costs involved for other types of NI are not available for the



Netherlands, but these are also expected to be substantial. A US study of 1355347 admissions in 55 US hospitals from 2001 to 2006 estimated that each nosocomial infection increased costs $12197 (Kilgore et al., 2008). A study for the Netherlands were only the prolonged length of stay in combination with infection control is estimated, calculated € 303 per hospital day in a ward and € 1141in an ICU. This gives a total cost estimation of € 2800 per NI patient. Table A.4: Prevalence studies of NIs in Dutch hospitals (Kamp-Hopmans et al.,

2003) (Hopmans et al., 2007) (van der Kooi et al., 2010)

Country The Netherlands The Netherlands The Netherlands Time span survey 1993-1998 2001-2004 2007-2008 No. of patients surveyed 3845 2661 26937 No. of hospitals surveyed 1 (UMC Utrecht) 2 (UMC Utrecht [n=2509]and

Central Military Hospital Utrecht[n=52])

41

Department/ward Surgery wards: Thoracic, gynaecologic, general and vascular, and orthopaedic

Departments: Surgery, neurology-neurosurgery, Heart-lung center, Internal medicine,

Otorhinolaryngology, Gynaecology, Paediatrics

Several

Patients with NI [-,%] 550 [14%] 295 [11.1%] 1667 [6.2%] No. of NIs [-] 648 340 1934 Mean LOS, patient with NI [days] 19.8 29.8 - Mean LOS, patient without NI [days]

7.7 17.5 -

Mortality due to NI [% in comparison to noninfected patients]

- 11.5% vs. 4.7 % -

Estimated extra costs due to NI [€ per NI patient]

- 2800 -

Mortality In addition to the increased hospitalization time, the contributing discomfort and increased economical cost, nosocomial infection also leads to mortality. Nosocomial infection mortality is an important problem, with estimates of magnitude placing it among the 10 leading causes of death in the United States (Haley, 1991). In 1988 9415 out of 217196 (0.43%) deaths in the USA had a nosocomial infection listed as contributing cause of death leading to an overall mortality rate of 3.83 per 100000 person-years (White, 1993). While persons dying with a nosocomial infection often have had a complicating condition and have received healthcare for an underlying disease, persons dying with a NI had a younger mean age of death and more causes listed on the death certificate as compared to deaths in the USA as a whole. However it is difficult to obtain a precise estimate of deaths from nosocomial infections, since patients often die from several causes and infection is often not mentioned on death certificates of patients who die from several causes and infection is often not mentioned on death certificates of patients who die of a combination of a chronic illness and acute infections (Curtis, 2008). A more recent study also based on patient records (Klevens et al., 2007), estimated a total number of deaths associated with an NI of nearly 99.000 in U.S. Hospitals in 2002. The number of deaths associated with an NI was highest for pneumonia (35%) and bloodstream infections (30%), an estimated 13% were associated with urinary tract infections and 8% with surgical site infections. Among the infected patients the highest percentage of patients whose death was associated with an infection was among adults and children in ICUs. The research done by Klevens and White reflects the difficulties to obtain precise number of deaths from NI’s, as there is a difference of tenfold between those studies. A Spanish prevalence study (Garcia-Martin et al., 2001) of 524 consecutive deaths in a Spanish 800-bed hospital supports the high number of NI related deaths, as 21.3% of the deaths of patients which occurred more than 48h after admission where due to nosocomial infections. A French study (Kaoutar et al., 2004) reported that 26.6% of the persons who died at least 48 hours after transmission, had an NI. NI contributed to the death in 14.6 % of the patients (certaine 6% and possible 8%). By the authors best knowing, there is no current data available for mortality in relation to NI in Dutch hospitals.



Reference list, Appendix A

Beggs, C. B. (2003). The airborne transmission of infection in hospital buildings: Fact or fiction? Indoor and Built Environment, 12, 9-18.

Caul, E. O. (1994). Small Round Structured Viruses - Airborne Transmission and Hospital Control. Lancet, 343, 1240-1242.

Christensen, M. & Jepsen, O. B. (2001). Reduced rates of hospital-acquired UTI in medical patients. Prevalence surveys indicate effect of active infection control programmes. Journal of Hospital Infection, 47, 36-40.

CIBSE, B., Barnard, N., & Jaunzens, D. (2004). CIBSE guide B: Section 2: Ventilation and air conditioning.

Curtis, L. T. (2008). Prevention of hospital-acquired infections: review of non-pharmacological interventions. Journal of Hospital Infection, 69, 204-219.

Erbaydar, S., Akgun, A., Eksik, A., Erbaydar, T., Bilge, O., & Bulut, A. (1995). Estimation of Increased Hospital Stay Due to Nosocomial Infections in Surgical Patients - Comparison of Matched Groups. Journal of Hospital Infection, 30, 149-154.

Eriksen, H. M., Iversen, B. G., & Aavitsland, P. (2005). Prevalence of nosocomial infections in hospitals in Norway, 2002 and 2003. Journal of Hospital Infection, 60, 40-45.

Franz, D. R., Jahrling, P. B., Friedlander, A. M., McClain, D. J., Hoover, D. L., Bryne, W. R. et al. (1997). Clinical recognition and management of patients exposed to biological warfare agents. Jama-Journal of the American Medical Association, 278, 399-411.

Garcia-Martin, M., Lardelli-Claret, P., Jimenez-Moleon, J. J., Bueno-Cavanillas, A., Luna-del-Castillo, J. D. D., & Galvez-Vargas, R. (2001). Proportion of hospital deaths potentially attributable to nosocomial infection. Infection Control and Hospital Epidemiology, 22, 708-714.

Gordts, B., Vrijens, F., Hulstaert, F., Devriese, S., & Van de Sande, S. (2010). The 2007 Belgian national prevalence survey for hospital-acquired infections. Journal of Hospital Infection, 75, 163-167.

Haley, R. W. (1991). Measuring the Costs of Nosocomial Infections - Methods for Estimating Economic Burden on the Hospital. American Journal of Medicine, 91, S32-S38.

Hopmans, T. E. M., Blok, H. E. M., Troelstra, A., & Bonten, M. J. M. (2007). Prevalence of hospital-acquired infections during successive surveillance surveys conducted at a university hospital in the Netherlands. Infection Control and Hospital Epidemiology, 28, 459-465.

Kamp-Hopmans, T. E. M., Blok, H. E. M., Troelstra, A., Gigengack-Baars, A. C. M., Weersink, A. J. L., Vandenbroucke-Grauls, C. M. J. E. et al. (2003). Surveillance for hospital-acquired infections on surgical wards in Dutch university hospital. Infection Control and Hospital Epidemiology, 24, 584-590.

Kaoutar, B., Joly, C., L'Heriteau, F., Barbut, F., Robert, J., Denis, M. et al. (2004). Nosocomial infections and hospital mortality: a multicentre epidemiological study. Journal of Hospital Infection, 58, 268-275.

Kilgore, M. L., Ghosh, K., Beavers, C. M., Wong, D. Y., Hymel, P. A., & Brossette, S. E. (2008). The costs of nosocomial infections. Medical Care, 46, 101-104.

Klevens, R. M., Edwards, J. R., Richards, C. L., Horan, T. C., Gaynes, R. P., Pollock, D. A. et al. (2007). Estimating health care-associated infections and deaths in US hospitals, 2002. Public Health Reports, 122, 160-166.

Kowalski, W. J. (2002). Immune building systems technology. New York: Mc Graw Hill.



Kowalski, W. J. (2009). Ultraviolet Germicidal Irradiation Handbook : UVGI for Air and Surface Disinfection. Berlin Heidelberg: Springer-Verlag Berlin Heidelberg.

RIVM, B. I. (2011). RIVM: Bulletins infectieziekten http://www.rivm.nl/Onderwerpen/Onderwerpen/I/Infectieziekten_Bulletin.

RIVM, M. (2008). RIVM; Melden van infectieziekten, conform de Wet publieke gezondheid (2008) http://www.rivm.nl/bibliotheek/rapporten/215072001.pdf.

Schaal, K. P. (1991). Medical and Microbiological Problems Arising from Airborne Infection in Hospitals. Journal of Hospital Infection, 18, 451-459.

Starakis, I., Marangos, M., Gikas, A., Pediaditis, I., & Bassaris, H. (2002). Repeated point prevalence survey of nosocomial infections in a Greek University hospital. Journal of Chemotherapy, 14, 272-278.

Tang, J. W., Li, Y., Eames, I., Chan, P. K. S., & Ridgway, G. L. (2006). Factors involved in the aerosol transmission of infection and control of ventilation in healthcare premises. Journal of Hospital Infection, 64, 100-114.

van der Kooi, T. I. I., Mannien, J., Wille, J. C., & van Benthem, B. H. B. (2010). Prevalence of nosocomial infections in The Netherlands, 2007-2008: results of the first four national studies. Journal of Hospital Infection, 75, 168-172.

Vaque, J., Rossello, J., & Arribas, J. L. (1999). Prevalence of nosocomial infections in Spain: EPINE study 1990-1997. Journal of Hospital Infection, 43, S105-S111.

White, M. C. (1993). Mortality Associated with Nosocomial Infections - Analysis of Multiple Cause-Of-Death Data. Journal of Clinical Epidemiology, 46, 95-100.

Xie, X. J., Li, Y. G., Sun, H. Q., & Liu, L. (2009). Exhaled droplets due to talking and coughing. Journal of the Royal Society Interface, 6, S703-S714.

http://www.rivm.nl/Onderwerpen/Onderwerpen/I/Infectieziekten_Bulletin�

http://www.rivm.nl/bibliotheek/rapporten/215072001.pdf�



Appendix B: Pictures measurement equipment

Figure B1: View on the measurement room, front manik in is the source patient,

manik in in the back is the rece iver patient .

Figure B.2: Left p ic ture: Supply gr id, posit ioned in the center of the ce i l ing. Right p icture: Exhaust gr id pos it ioned above the door

Figure B.3: Outs ide of the test room, with the balance venti lat ion system at the forefront.



(a) (b) Figure B.4: (a) Source manikin with the breathing dev ice at h is mouth. (b) Receiver manik in



Appendix C: Measurement equipment The measurement equipment used in this study is all property of the building physics and services laboratory at the Eindhoven University of Technology. An overview of the used sensors with their accuracies is given in Table C.1 Table C.10: Used measurement equipment. Variable Type sensor TU/e ID

no. Output Range Accuracy

Toutside U Thermistor - -50 to 150 ℃ -50 to 150 ℃ Unknown Troom U Thermistor - -50 to 150 ℃ -50 to 150 ℃ Unknown Tsupply air U Thermistor - -50 to 150 ℃ -50 to 150 ℃ Unknown CO2-Room Bruel & Kjaer multi-gas

monitor (type 1302), with a multi-point sampler (type 1303)

320 ppm CO2 0->10% 1.7 ppm

CO2-Supply air Vaisala CaTec C02 sensor 328 0 – 10 [V] 0-2000ppm 40 ppm CO2-Exhaust air Vaisala CaTec CO2 sensor 341 0 – 10 [V] 0-2000ppm 40 ppm CO2-Roomcontrol Vaisala CaTec CO2 sensor 329 0 – 10 [V] 0-5000ppm 100ppm �̇�C02-exhalation Read out Brooks Instrument

0152 & Mass flow controller Brooks Instrument 5850S

0807/0808 % of 15L 0-15L 0.015L

�̇�air-exhalation 2x flowmeter (parallel) 1863/1864 0-2.5 and 0-5 L/min

0-2.5 and 0-5 L/min 0.2 l/min

�̇�supply air pillow Kalinsky sensor elektronik measurement pipe D9669 with Pressure Gauge HMG 6-1

1180/1982 0-320Pa 30-450 m3/hr 2 m3 /hr

�̇�supply air room

ventilation

Acin Flow finder 2289 2-300 m3/hr 2- 300 m3/hr 3.5% of reading

�̇�exhaust air room

ventilation

Acin Flow finder 2289 2-300 m3/hr 2-300 m3/hr 3.5% of reading

Psource manikin Voltcraft Energy Logger 4000 2268 Watts unknown ±1.5 W Preceiver manikin Voltcraft Energy Logger 4000 2090 Watts unknown ±1.5 W Movement person Resistance 2-wire - 0-30000R 0-30000R unknown nparticle-patient MetOne Remote Particle

counter R4903 0083 >0.3𝜇𝑚 = 0-5

[V] >0.5𝜇𝑚 = 0-5 [V]

with PV: >0.3𝜇𝑚 = 0-5000 >0.5𝜇𝑚 = 0-1000 without PV: >0.3𝜇𝑚 = 0-25000 >0.5𝜇𝑚 = 0-5000

5 % coincidence loss

nparticle-outside MetOne Remote Particle counter R4903

0239 >0.3𝜇𝑚 = 0-5 [V] >0.5𝜇𝑚 = 0-5 [V]

>0.3𝜇𝑚 = 0-6000 >0.5𝜇𝑚 = 0-1500

5%

nparticle-exhaust MetOne Remote Particle counter R4903

>0.3𝜇𝑚 = 0-5 [V] >0.5𝜇𝑚 = 0-5 [V]

>0.3𝜇𝑚 = 0-400000 >0.5𝜇𝑚 = 0-12000

5%

nparticle-control Particele counter Ligthhouse R 2014

1840 particles 0-2000000 particles/ft3

5 % coincidence loss

Air velocity Testo 425 217 m/s 0-20 m/s ±(0.03 m/s +5% of meas. value

Sound level Bruel & Kjaer 2232 983 dB 34-140dB ±1 dB



Appendix D: Measurement protocol. The experiments are numbered according to the following Table. Also smoke gas visualizations are made of all movements. A1 Validating of the mixing-ventilation pattern in the test room. With smoke gas A2 Normal CO2 concentrations Running the equipment for one day, no breathing, no personal ventilation, Qin=112m3/h B “no movements” -with CO2 tracer gas Test Orientation

manikins Exhalation mode Personal ventilation

source patient Personal ventilation receiver patient

Movements

B1 Face up Constant Breathing no no no B2 Face up Constant Breathing yes yes no B3 Face up Constant Breathing yes no no B4 Face up Constant Breathing no yes no B5 Face to face Constant Breathing no no no B6 Face to face Constant Breathing yes yes no B7 Face to face Constant Breathing yes no no B8 Face to face Constant Breathing no yes no C Continuous movement of a person in between the beds - with CO2 tracer gas Test Orientation



Movements

C1 Face up Constant Breathing no no Moving person C2 Face up Constant Breathing yes yes Moving person C3 Face up Constant Breathing yes no Moving person C4 Face up Constant Breathing no yes Moving person C5 Face to face Constant Breathing no no Moving person C6 Face to face Constant Breathing yes yes Moving person C7 Face to face Constant Breathing yes no Moving person C8 Face to face Constant Breathing no yes Moving person D Moving person (1 movement) in between the beds -with smoke gas and particle counts analyzing Test Orientation



Movements

D1 Face to face Constant Breathing no no Moving person D2 Face to face Constant Breathing yes yes Moving person E Moving bed-sheets -with smoke gas and particle counts analyzing Test Orientation



Movements

E1 Face to face Constant Breathing no no Moving bed-sheets E2 Face to face Constant Breathing yes yes Moving bed-sheets F Swinging door - with smoke gas and particle counts analyzing Test Orientation



Movements

F5 Face to face Constant Breathing no no Swinging door F6 Face to face Constant Breathing yes yes Swinging door Planning Measurements will be done in alphabetical order. First A, then the whole set of B, etc. Starting up the experiments till all temperatures and concentrations have reached steady state takes 45 minutes (figure D.1).



These measurements have taken one month. In a working day of 8 hours, 2 series of more than 3 hours could be done, so each test consists of 36-40 measurement points per position. 2 tests C4 and B6 are repeated. Due to the sub-conclusions of the CO2 concentration experiments; only face to face experiments in the both PV on and PV off configuration are tested for the single time dynamic measurements. In two weeks there is time for 60 measurements in 2 sets of 30 per serie. Protocol Before the start of a measurement.

-Heating of the thermal manikins, this process takes 30 minutes until a steady-state is reached (Figure D.). The power input of both manikins will be checked before the start of each experiment.

-CO2 distribution and breathing mechanism will be started, 30 minutes before the start of a measurement. The measurement will only start when steady state of the CO2 concentration in the exhaust air is reached.

-Lights will be set on, 30 minutes before the start. -Cooling of the supply air will be set on, 30 minutes before the start. Measurement will only start when

supply air temperature has reached a steady-state. -Room ventilation will be activated 30 minutes before the start of each experiment. -The breathable pillow(s) will be activated 30 minutes before the start of the measurements in case of

the personal ventilation experiments. -The moving manikin will be activated 30 minutes before the start of the measurements in case of the

experiments with movements. -There room will not be entered from 30 minutes before the start of the experiment, till the end of the

experiment. During the experiments

-Logging of temperature, supply air CO2 concentration and exhaust air CO2 concentration every 10 seconds. -Logging of movement of person, every second. -Logging of CO2 concentration inside the room approximately every minute. -Visual control every 15 minutes of:

• Lights • Air flow rate of breathing mechanism with CO2 supply system. • Supply rate of CO2 • Working of room ventilation system • Air flow rate of personal ventilation system • Room temperature • Supply air temperature • Supply and exhaust CO2 concentration • Working of moving manikin • Working of the multi-gas analyzer

When one of these parameters is wrong, the measurements will be restarted. Immediately after the measurements Visual Control of :

• Power input of both thermal manikins • Lights • Air flow rate of breathing mechanism with CO2 supply system. • Supply rate of CO2 • Working of room ventilation system



• Air flow rate of personal ventilation system • Room temperature • Supply air temperature • Supply and exhaust CO2 concentration • Working of moving manikin • Working of the multi-gas analyzer

Figure D.1: Surface temperature of the legs and body of the receiver patient by warming up.

0

5

10

15

20

25

30

35

40

0 5 10 15 20 25 30 35 40 45 50 55 60 65

Tem

pera

ture

[C]

Time [min]

Right leg

Left leg

Body

environment



Appendix E: Calibration Calibration CO2 sensors.

Figure E.1: Measurement set-up of the calibration. CO2 concentration is slightly decreased during time and measured by all sensors. A calibration of the CO2 sensors has been made 2 year before the measurements and is repeated right after the measurements. Calibration is done by placing all CO2 sensors in box and measuring the decrease of CO2 concentration in that box (Figure E.2), the Bruel and Kjaer is used as a reference during these experiments. Results of a comparison between those calibration (Figure E.6 and E.7) demonstrate that the sensors have remain constant during this period. Also a control measurement at the start of the measurement series (See Appendix H, measurement A.2) shows that the CO2 sensors are comparable to each other. A Matlab M-file is used to process the calibration results.

figure plot(A4_ch1(80:1825,5),A4_ch1(80:1825,3),'.') xlabel('[V] sensor 328') ylabel('[ppm] B&K') fit=polyfit(A4_ch1(200:1825,5),A4_ch1(200:1825,3),2); f=polyval(fit,A4_ch1(200:1825,5)); bk=A4_ch1(200:1825,3); error=f-bk; figure plot(error) ylabel('error [ppm]') xlabel('measurement no.') gem=mean(abs(error))

Figure E.2: Scatter graph of calibration of sensor 328 to the Bruel & Kjaer



The polyfit of the supply air CO2 sensor , ID 328 is: -1.2148x2 + 189.8850x + 100.5975 The mean absolute error of the supply air CO2 sensor 328 in relation to this polyfit is (figure E.3): 4.67 ppm

Figure E.3: Error graph of sensor 328 Sensor 341:

Figure E.4: Scatter graph of calibration of sensor 341 in relation to the Bruel & Kjaer The polyfit of the exhaust air CO2 sensor , ID 341 is: -1.5373x2 + 191.6807x + 66.4161 The mean absolute error of the exhaust air CO2 sensor 341 in relation to this polyfit is (figure E.5): 4.92 ppm



Figure E.5: Error graph of sensor 341 With Matlab a comparison with earlier calibration is made. The results are comparable and therefore it can be assumed that the sensors have remain constant during all tests. :

x=1.5:0.1:5; %y2=-1.5373* x.^2 + 191.6807* x +66.4164; y2=-1.5835* x.^2 + 192.3676* x +76.7928; y1=0.0994* x.^2 + 181.7* x + 98.965; %y1=-0.2239* x.^2 + 175.41* x + 94.169; error=y2-y1; plot(x,y2) hold on plot(x,y1,'r') legend('calibration 17 April 2012','calibration 15 March 2010','Location','Southeast') xlabel('Sensor 328 [V]') Ylabel('ppm') Title('Comparison of calibrations')

Figure E.6: Comparison of two calibrations of the supply air sensor



Figure E.7: Comparison of two calibrations of the exhaust air sensor



Calibration particle counters A Calibration of the particle counters is made before all experiments and after all experiments. A calibration is done by distributing smoke in a closed box and measurement of the decrease of particles as function of time. Results of the particles sized 0.3 𝜇m to and particles larger than 0.5 𝜇m are shown in figure E.8 and E.9 respectively. As there was not a calibrated sensor available, results are compared to a particle counter of another type (Lighthouse R 2014) which was not used during the experiments, and therefore acts as the control counter in this calibration. As this sensor only counts particles larger than 0.5 𝜇m, real values of the particle counters are not available and only relative numbers can be used. It can be seen in Figure E.8 that the smaller particle sensors remained constant, as no difference of voltage, which has a linear relation to particle counts can be seen. In figure E.9 can be seen that the exhaust and room particle sensor of particles larger than 0.5 𝜇m remains constant, and therefore results of these sensors are reliable, the outside sensor results differ a factor 1.5 and therefore not reliable.

Figure E.8: Calibration of the 0.3𝜇m particle counters used in the experiments to the reference counter, the 0.5𝜇m Lighthouse R2014; 14May=date before all experiments, 11June=date after all experiments.

Figure E.9: Calibration of the 0.5𝜇m particle counters used in the experiments to the reference counter, the 0.5𝜇m Lighthouse R2014; 14May=date before all experiments, 11June=date after all experiments.



Appendix F: Conversation hospital ventilation 8-6-2012 Samenvatting van een gesprek met Albert Trip, manager nieuwbouw Meander Medisch Centrum en Reinder Bielleman Projectmanager en installatiekostendeskundige Lisoba Beheer bv “100% van de toekomst van ziekenhuizen ligt in het voorkomen van besmettingen.” Opbouw nieuwe Meander Ziekenhuis: Gekozen voor 100% eenpersoonskamers met eigen sanitair (Orbis Medisch Centrum in Sittard-Geleen is het eerste en enig andere ziekenhuis in Nederland). Van de eenpersoonskamers zijn er 8 kamers per 96 gesluisd. Meeste ziekenhuizen kiezen voor een mix van 50% van eenpersoonskamers en meerpersoonskamers. Meander heeft gekozen voor eenpersoonskamers om het infectierisico te verminderen en wil de patiënten ook meer privacy geven. De kamers zijn ingericht volgens het Healing Environment principe, wat inhoudt dat een patiënt zich op zijn gemak voelt, snel het genezingsproces kan inzetten en daarna thuis verder herstelt. Het ziekenhuis moet dan ook niet gezien worden als een plaats waar de patiënt van begin tot het eind van zijn ziekteverschijnselen verblijft, maar als een soort hotel waar de patiënt alleen voor de meest noodzakelijk zorg verblijft. Een zo kort mogelijk verblijf is dan ook wenselijk met het oog op herstel van de patiënt en ter voorkoming van besmetting en infecties. Om patiënten niet te laten vereenzamen, wordt verblijf van een familielid op de kamer mogelijk gemaakt. De grote lage ramen maken visueel contact met de buitenwereld mogelijk. De sanitaire ruimtes liggen tussen de kamers in, waardoor de kamer de volle breedte houdt van gevel tot gang. De brede schuifdeur geeft toegang tot het verblijfsgebied buiten de kamer, waar de lounge mogelijkheden biedt voor de reactivering van de patiënt ter bevordering van zijn herstelproces. De kamers hebben een toevoerrooster boven de deur en een afvoerrooster in de sanitaire ruimtes. Daarnaast kunnen op vraag van de patiënt beide kantelramen opengezet worden. De lucht wordt bevochtigd met stoom. De kamers worden geventileerd met 80m3/hr op een 3 meter hoge kamer van 15m2 vloeroppervlakte. Dit is een lage ventilatievoud van 1.8 ACH wat maar net boven de eisen van het bouwbesluit is. De filosofie hierachter is dat meer lucht ook zorgt voor meer luchtbeweging. Samen met betonkernactivering en het niet gebruik maken van radiatoren, verwacht het ziekenhuis zo een zo laminair mogelijke stroming. Andere ziekenhuizen kiezen voor hogere ventilatievouden (Deventer 4 ACH) maar met het oog op duurzaamheid en kostenbesparing levert een hogere ventilatievoud al snel een hogere energiebelasting op. In het Meander ziekenhuis bedraagt de totale ventilatie 500000 m3/hr en levert iedere m3/hr minder lucht een besparing op van 25 euro op jaarbasis. Infectierisico Het overgrote deel van de patiënten is niet verminderd vatbaar door de aandoening waarvoor ze in het ziekenhuis behandeld worden. Weerstand tegen ziektes kweek je door licht aan de ziektekiemen blootgesteld te worden, een totale isolatie van ziektekiemen is voor veel patiënten dan ook niet noodzakelijk. Echter voor oncologische patiënten, patiënten met een verminderd immuunsysteem bijvoorbeeld door HIV of internistiche patiënten die juist van een operatiekamer afkomen is de vatbaarheid voor ziektes hoog en het infectierisico moet dan ook zeker voor deze patiënten geminimaliseerd worden. Deze patiëntengroepen representeren ongeveer 20 % van de totale bevolking in patiëntenkamers. De leukopene patiënten worden verzorgd in 4 speciale kamers met overdruk. Voor de meeste patiënten en medewerkers geldt dat ze niet ziek worden van kortstondige blootstelling aan een geïnfecteerde patiënt, maar dat langdurige blootstelling het probleem is. Het infectierisico wordt dan ook maar minimaal verhoogd als een patiënt op een bed van de ene plaats naar de andere wordt gereden langs andere patiënten. Er ontstaat pas serieus infectierisico bij blootstelling van tenminste 15 minuten. De filosofie in het Meander is om de infectiebron te isoleren en niet zozeer om elke patiënt afzonderlijk te isoleren, daarom is er een lichte onderdruk in de kamers, zodat de omgeving “schoon” blijft.



Nut van Persoonlijke ventilatie Persoonlijke ventilatie zou nuttig kunnen zijn doordat het een microklimaat kan bezorgen bij de kwetsbare patiënt met verhoogd risico bij infecties. Bij patiënten die een langdurige operatie hebben ondergaan is afkoeling en onderkoeling een gevaarlijk probleem. Bij narcosepatiënten is het van belang dat ze zo snel mogelijk uit hun narcose komen maar bij slechte ventilatie rondom het gezicht blijven zij hun uitgeademde narcosegassen inademen, waardoor ze langzaam ontwaken. PV kan dit proces versnellen door extra toevoer van verse lucht of door lage afzuiging van narcosegassen (toepassing op de verkoeverruimte). Ook bij patiënten die te kampen hebben met doorliggen (Decubitus) zou PV een goede oplossing kunnen zijn. Nadeel van PV is dat een bed in een patiëntenkamer mobiel is en veel PV systemen daar moeilijk op aan te passen zijn. Ook zijn nog meer systemen niet wenselijk rondom een bed. Vier infuuspompen rondom een bed zijn niet uitzonderlijk. De verpleging moet ongehinderd zijn werk kunnen doen en zo min mogelijk overbodige handelingen moeten verrichten.. Beweging In beddenhuizen zijn er 6 verpleegkundigen per 30 patiënten continu bezig met handelingen rond de patiënt. Daarnaast zijn er facilitair medewerkers, artsen en bezoekers die ook in de patiëntenkamer komen. Op intensieve zorg afdelingen is er gemiddeld 1 verpleegkunige per 2 patiënten aanwezig in de patiëntenkamers. Dit betekent niet dat er de helft van de tijd mensen bij de patiënt zijn want op sommige tijden staan er 4 medewerkers bij een patiënt, maar hier wordt minder bezoek toegelaten. Patiënten worden aangemoedigd om zoveel mogelijk te bewegen om hun spiermassa niet af te laten nemen en dus zo fit mogelijk het ziekenhuis te kunnen verlaten.

Figure J.1: Plan of a standard patient room floor at Meander Medisch centrum

Figure J.2: Render of a new en-suite one person patient room in Meander Medisch Centrum.



Appendix G: Results; performance of PV when only one pillow is active

Figure G.1: Position of the measurement points to measure CO2 concentration

Static experiments Table G.1: Local air quality [LAQ] index at different positions (Figure 7) for the static experiments.

Position 1 LAQ [-]

Position 2 LAQ [-]

Position 3 LAQ [-]

Position 4 LAQ [-]

Position 5 LAQ [-]

Patients lying both face up PV off 0.94 0.96 0.97 0.97 0.98 PV on at both patients 0.58 1.35 1.35 1.53 1.72 PV on only at source patient 0.62 1.15 1.16 1.18 1.23 PV on only at receiver patient 1.03 1.17 1.15 1.46 1.75 Patients lying face to face PV off 0.75 0.91 0.83 0.97 0.98 PV on at both patients 0.99 0.84 0.90 0.99 1.64 PV on only at source patient 0.77 0.63 0.70 0.81 0.85 PV on only at receiver patient 0.77 0.63 0.80 1.36 1.67

Table G.2: Effectiveness of personal ventilation system when PV is only active at the source patient; at different positions (Figure 7) for the static experiments. A positive value (max. 1=100% clean air) means an improvement, a negative value means a worsening of the local air quality due to the implementation of PV system for both patients, see Chapter 3.

Position Effectiveness of PV for PV only at the source patient in comparison

with no PV [-]

Effectiveness of PV for PV only at the source patient in comparison

with no PV [-] Patients lying both face up Patients lying face to face

1. 0.5m above source patient -1.09 0.04 2. Middle, 1.2m above floor 0.36 -0.97 3. Middle, 1.65m above floor 0.37 -0.37 4. 0.5m above receiver patient 0.41 -0.47 5. At receiver patients’ mouth 0.45 -0.37

Table G.3: Effectiveness of personal ventilation system when PV is only active at the receiver patient; at different positions (Figure 7) for the static experiments. A positive value (max. 1=100% clean air) means an improvement, a negative value means a worsening of the local air quality due to the implementation of PV system for both patients, see Chapter 3.


with no PV [-]



1. 0.5m above source patient 0.19 0.05 2. Middle, 1.2m above floor 0.38 -0.94 3. Middle, 1.65m above floor 0.35 -0.08 4. 0.5m above receiver patient 0.75 0.68 5. At receiver patients’ mouth 0.98 0.98



Continuous movements Table G.4: Local air quality [LAQ] index at different positions (Figure 7) for the continuous movement experiments.

Position 1 LAQ [-]

Position 2 LAQ [-]

Position 3 LAQ [-]

Position 4 LAQ [-]

Position 5 LAQ [-]

Patients lying both face up PV off 0.91 0.94 0.95 0.96 0.96 PV on at both patients 1.00 1.13 1.14 1.25 1.56 PV on only at source patient 0.84 0.99 1.01 1.01 1.01 PV on only at receiver patient 0.86 0.93 0.94 1.08 1.57 Patients lying face to face PV off 0.80 0.90 0.89 0.95 0.97 PV on at both patients 0.97 0.81 0.85 0.95 1.58 PV on only at source patient 0.97 0.85 0.83 0.86 0.86 PV on only at receiver patient 0.89 0.91 0.89 1.20 1.61

Table G.5: Effectiveness of personal ventilation system when PV is only active at the source patient; at different positions (Figure 7) for the continuous movement experiments. A positive value (max. 1=100% clean air) means an improvement, a negative value means a worsening of the local air quality due to the implementation of PV system for both patients, see Chapter 3.


with no PV [-]



1. 0.5m above source patient -0.19 -0.14 2. Middle, 1.2m above floor 0.11 -0.02 3. Middle, 1.65m above floor 0.12 -0.04 4. 0.5m above receiver patient 0.12 0.25 5. At receiver patients’ mouth 0.11 0.89

Table G.6: Effectiveness of personal ventilation system when PV is only active at the receiver patient; at different positions (Figure 7) for the continuous movement experiments. A positive value (max. 1=100% clean air) means an improvement, a negative value means a worsening of the local air quality due to the implementation of PV system for both patients, see Chapter 3.


with no PV [-]



1. 0.5m above source patient 0.34 0.21 2. Middle, 1.2m above floor -0.14 0.02 3. Middle, 1.65m above floor -0.15 -0.01 4. 0.5m above receiver patient -0.26 0.45 5. At receiver patients’ mouth -0.32 0.90



Appendix H: Results graphs This appendix shows graphs of the raw data and the boundary conditions. Experiment numbering is according to the table in Appendix D. For series A, B and C the blue line in every graph represents the measured value at a specific point in the room (Table H.1). The green and red line are supply and exhaust conditions measured at the same time as the blue line. Table H.1: Position of sensors and position of graphs Position sensor in room Annotation in

graph Position of graph in sub-plots

1. 0.5m above source patient ch1 Upper row, first picture 2. Middle, 1.2m above floor ch2 Upper row, second picture 3. Middle, 1.65m above floor ch3 Upper row, third picture 4. 0.5m above receiver patient ch4 Lower row, first picture 5. At receiver patients’ mouth ch4 Lower row, second picture For the single time movement series D, E and F the only measurement position in the room is at the receiver patients’ mouth. Outside particle counts are done just outside the test room, so inside the building physics and services laboratory and the sensor is positioned 1m above floor level, and not interfering with the exhaust jet. Exhaust particle counts are taken at the end of the exhaust tube. For the case where the person was moving and PV was on (2 series) and 1 series when PV was off, no measurements are done at this position. The green line in the main graph represents the movement; when this line is changing from 0 to maximum, then the movement is started, when this value changes from maximum to zero, the return movement is working.



A2: Steady state conditions

B1: Static situation; both patients lying face up; No PV



B2: Static situation; both patients lying face up; both PV active

B3: Static situation; both patients lying face up; only source patient’s PV active



B4: Static situation; both patients lying face up; only receiver patient’s PV active

B5: Static situation; both patients lying face to face; No PV



B6: Static situation; patients lying face to face; both PV active

B6_2: Static situation; patients lying face to face; both PV active; control measurement



B7: Static situation; patients lying face to face; only source patient’s PV active

B8: Static situation; patients lying face to face; only receiver patient’s PV active



C1: Continuous dynamic situation; both patients lying face up; No PV

C2: Continuous dynamic situation; both patients lying face up; both PV active



C3: Continuous dynamic situation; both patients lying face up; only source patient’s PV active

C4: Continuous dynamic situation; both patients lying face up; only receiver patient’s PV active



C4_2: Continuous dynamic situation; both patients lying face up; only receiver patient’s PV active, control measurement

C5: Continuous dynamic situation; both patients lying face to face; No PV



C6: Continuous dynamic situation; patients lying face to face; both PV active

C7: Continuous dynamic situation; patients lying face to face; only source patient’s PV active



C8: Continuous dynamic situation; patients lying face to face; only receiver patient’s PV active



D Single time moving person, PV off, 1st set of measurements



D Single time moving person, PV off, 1st set of measurements Boundary conditions



D Single time moving person, PV off, 2nd set of measurements



D Single time moving person, PV off, 2nd set of measurements Boundary conditions



D Single time moving person, PV on, 1st set of measurements



D Single time moving person, PV on, 1st set of measurements Boundary conditions



D Single time moving person, PV on, 2nd set of measurements



D Single time moving person, PV on, 2nd set of measurements Boundary conditions



E Single time turning over of bed-sheets, PV off, 1st set of measurements



E Single time turning over of bed-sheets, PV off, 1st set of measurements Boundary conditions



E Single time turning over of bed-sheets, PV off, 2nd set of measurements



E Single time turning over of bed-sheets, PV off, 2nd set of measurements Boundary conditions



E Single time turning over of bed-sheets, PV on, 1st set of measurements



E Single time turning over of bed-sheets, PV on, 1st set of measurements Boundary conditions



E Single time turning over of bed-sheets, PV on, 2nd set of measurements



E Single time turning over of bed-sheets, PV on, 2nd set of measurements Boundary conditions



F Single time opening and immediately closing of a door, PV off, 1st set of measurements



F Single time opening and immediately closing of a door, PV off, 1st set of measurements Boundary conditions



F Single time opening and immediately closing of a door, PV off, 2nd set of measurements



F Single time opening and immediately closing of a door, PV off, 2nd set of measurements Boundary conditions



F Single time opening and immediately closing of a door, PV on, 1st set of measurements



F Single time opening and immediately closing of a door, PV on, 1st set of measurements Boundary conditions



F Single time opening and immediately closing of a door, PV on, 2nd set of measurements



F Single time opening and immediately closing of a door, PV on, 2nd set of measurements Boundary conditions



Appendix I: Statistical analysis Statistical analysis of static and continuous measurements. A mean, median, mode and standard deviation of the measured CO2 concentration in the room is described in the tables below. The data is also analyzed on a normal distribution with the Lilliefors test an adaptation of the Kolmogorov-Smirnov test. Lilliefors test =𝑀𝑎𝑥𝑥|𝑆𝐶𝐷𝐹(𝑥) − 𝐶𝐷𝐹(𝑥)| Where, SCDF is the empirical cdf estimated from the sample

CDF is the normal cdf with mean and standard deviation equal to the mean and standard deviation of the sample.

Lilliefors test in Matlab: Lillietest uses a table of critical values computed using Monte Carlo simulation for sample sizes less than 1000 and significance levels between 0.001 and 0.50. The table is larger and more accurate than the table introduced by Lilliefors. Critical values for a test are computed by interpolating into the table, using an analytic approximation when extrapolating for larger sample sizes [Matlab 2010 B]. Lilliefors uses a null hypothesis that the sample in vector x comes from a distribution in the normal family, against the alternative that it does not come from a normal distribution. The test returns the logical value h = 1 if it rejects the null hypothesis at the 5% significance level, so then the distribution is not normally distributed, and h = 0 if it cannot reject the null-hypothesis [Matlab 2010B]. The test series are numbered according to the schedule in Appendix D. The control point sensor is a Vaisala Catec CO2 sensor positioned nearby position 1 of the B&K. Measurement B6 and C4 have been done twice. Static measurements; patients lying both face up B1

B&K Supply Exhaust Control point

Kanaal mean std median mode lilliefors p (lillie) mean std mean std mean std

1 845 109 799 784 1 0.00 436 6 750 14 762 87

2 786 21 781 798 0 0.15 437 5 752 17 0 0

3 782 23 775 766 1 0.05 437 6 752 18 0 0

4 775 12 775 779 0 0.47 436 6 749 20 0 0

5 767 14 765 765 0 0.19 435 5 751 14 0 0

B2



1 1459 584 1425 1210 0 0.50 470 13 825 31 742 433

2 637 61 613 581 1 0.00 470 13 827 28 0 0

3 642 67 618 600 1 0.00 471 13 832 30 0 0

4 553 40 543 532 1 0.01 470 12 830 30 0 0

5 478 11 478 465 0 0.50 471 11 823 23 0 0



B3



1 1968 1625 1330 1330 1 0.00 468 7 825 28 1033 995

2 722 61 710 718 1 0.00 468 6 814 31 0 0

3 718 63 701 670 1 0.02 469 6 814 28 0 0

4 694 48 692 664 1 0.01 469 7 815 27 0 0

5 681 39 671 660 1 0.04 468 6 824 28 0 0

B4



1 1132 661 792 732 1 0.00 464 9 814 25 -126 0

2 717 69 703 697 1 0.00 462 7 819 24 0 0

3 717 69 706 648 1 0.00 463 8 812 26 0 0

4 574 48 561 511 1 0.01 463 7 822 28 0 0

5 469 12 470 467 0 0.35 463 6 822 28 0 0

Static measurements; patients lying face to face B5



1 1034 155 1040 1010 0 0.50 462 8 783 15 855 95

2 956 193 864 1100 1 0.00 461 9 786 16 0 0

3 989 186 943 815 1 0.01 462 9 783 16 0 0

4 821 68 810 797 1 0.00 461 6 784 13 0 0

5 803 24 804 806 0 0.06 461 8 785 15 0 0

B6



1 747 112 759 774 1 0.04 443 6 750 9 498 65

2 892 80 896 796 0 0.22 442 5 754 10 0 0

3 830 31 834 779 0 0.38 442 7 753 9 0 0

4 757 68 760 699 0 0.35 442 7 751 9 0 0

5 459 5 458 461 0 0.20 441 6 751 10 0 0



B6_2



1 672 67 683 667 0 0.18 430 5 728 7 526 45

2 898 77 890 839 0 0.27 431 5 728 9 0 0

3 837 47 827 814 1 0.02 429 7 729 10 0 0

4 725 62 720 676 0 0.50 429 6 728 11 0 0

5 451 15 457 458 1 0.00 429 5 730 8 0 0

B7



1 971 202 970 1030 0 0.16 447 8 750 11 463 42

2 1295 309 1190 1120 1 0.01 447 8 750 12 0 0

3 1104 185 1070 1130 0 0.23 448 9 751 13 0 0

4 956 117 926 1020 1 0.03 448 10 749 11 0 0

5 918 99 886 814 1 0.02 447 10 750 11 0 0

B8



1 1015 174 1030 1190 0 0.50 467 11 794 22 853 89

2 1220 360 1250 982 0 0.50 466 11 791 22 0 0

3 1068 238 991 941 1 0.00 466 9 791 22 0 0

4 610 78 581 531 1 0.01 468 10 789 16 0 0

5 475 22 476 458 0 0.26 468 12 795 21 0 0

Continuous dynamic measurements; patients lying both face up C1



1 1006 299 879 874 1 0.00 457 7 798 12 716 63

2 847 22 845 814 0 0.50 458 8 796 13 0 0

3 843 21 838 831 0 0.24 458 7 797 11 0 0

4 831 19 828 852 0 0.20 459 7 796 12 0 0

5 827 18 829 812 0 0.50 458 6 797 11 0 0

C2



1 1047 551 792 683 1 0.00 473 9 791 18 640 97

2 707 39 701 643 0 0.17 474 9 789 15 0 0

3 692 30 691 682 0 0.43 475 8 790 15 0 0

4 644 42 632 631 0 0.06 473 9 787 14 0 0

5 509 38 506 495 1 0.00 474 8 788 17 0 0



C3



1 1162 480 940 1140 1 0.00 459 17 785 27 877 259

2 796 34 790 771 0 0.33 459 15 783 22 0 0

3 782 25 779 763 0 0.13 461 17 783 22 0 0

4 775 26 771 716 0 0.16 461 15 783 25 0 0

5 771 24 774 739 0 0.15 460 15 782 23 0 0

C4



1 1023 297 886 816 1 0.00 446 8 758 24 1023 272

2 819 36 818 784 0 0.24 445 9 764 32 0 0

3 822 51 814 781 0 0.08 446 10 761 26 0 0

4 708 53 704 684 0 0.50 446 10 760 26 0 0

5 497 35 485 481 1 0.00 446 8 762 29 0 0

C4_2



1 928 300 800 793 1 0.00 433 5 747 11 695 64

2 771 36 774 771 0 0.10 434 8 749 14 0 0

3 766 30 767 784 0 0.37 434 5 748 13 0 0

4 663 56 659 634 0 0.22 433 5 749 11 0 0

5 500 37 497 490 0 0.50 433 7 748 12 0 0

Continuous dynamic measurements; patients lying face to face C5



1 989 125 966 831 1 0.01 449 5 770 8 773 53

2 965 288 853 805 1 0.00 449 6 771 7 0 0

3 893 81 863 852 1 0.01 449 6 768 9 0 0

4 812 24 808 797 0 0.16 450 5 770 8 0 0

5 801 30 794 777 1 0.00 449 5 770 8 0 0



C6



1 790 86 782 615 0 0.50 450 7 755 14 571 56

2 937 56 931 908 0 0.11 449 7 754 14 0 0

3 901 49 893 887 1 0.02 448 6 755 13 0 0

4 790 80 797 820 0 0.50 449 6 756 13 0 0

5 476 19 478 474 1 0.01 448 9 756 13 0 0

C7



1 780 134 785 749 0 0.19 461 8 764 7 513 32

2 935 164 896 880 1 0.00 460 7 766 8 0 0

3 940 67 919 875 1 0.01 461 6 767 8 0 0

4 895 39 891 903 0 0.50 462 6 766 8 0 0

5 890 50 894 857 0 0.50 461 7 766 7 0 0

C8



1 841 115 830 721 0 0.24 429 5 743 19 756 62

2 845 99 819 848 0 0.06 430 5 745 16 0 0

3 845 66 841 778 0 0.13 430 6 745 16 0 0

4 651 95 622 572 1 0.00 429 4 745 17 0 0

5 460 17 463 463 1 0.03 429 6 745 18 0 0



Statistical analysis of exhaust and outside particle counters DPVaan3 Mean Standard

Deviation Exhaust 0.3-0.5 𝝁m - - Exhaust >0.5 𝝁m - - Outside 0.3-0.5 𝝁m 146.782 107.9855 Outside >0.5 𝝁m 5.875279 9.632241 DPVuit5 Mean Standard

Deviation Exhaust 0.3-0.5 𝝁m - - Exhaust >0.5 𝝁m - - Outside 0.3-0.5 𝝁m 182.9705 74.89667 Outside >0.5 𝝁m 1.372671 0.469486 EPVaan2 Mean Standard

Deviation Exhaust 0.3-0.5 𝝁m 4026.323 276.5138 Exhaust >0.5 𝝁m 11.84485 6.750174 Outside 0.3-0.5 𝝁m 260.6006 27.1543 Outside >0.5 𝝁m 0.96422 0.350582 EPVuit1 Mean Standard

Deviation Exhaust 0.3-0.5 𝝁m 5005.465 466.9936 Exhaust >0.5 𝝁m 28.61442 12.33186 Outside 0.3-0.5 𝝁m 300.5147 34.22996 Outside >0.5 𝝁m 1.037386 0.742435 FPVaan1 Mean Standard

Deviation Exhaust 0.3-0.5 𝝁m 5652.783 797.9423 Exhaust >0.5 𝝁m 13.67268 13.77301 Outside 0.3-0.5 𝝁m 310.9062 54.57301 Outside >0.5 𝝁m 7.624834 7.317877 FPVuit2 Mean Standard

Deviation Exhaust 0.3-0.5 𝝁m 5634.308 238.2019 Exhaust >0.5 𝝁m 37.28123 11.63249 Outside 0.3-0.5 𝝁m 212.2794 36.82873 Outside >0.5 𝝁m 1.968721 0.911593

DPVaan4 Mean Standard Deviation

Exhaust 0.3-0.5 𝝁m - - Exhaust >0.5 𝝁m - - Outside 0.3-0.5 𝝁m 200.8723 89.7132 Outside >0.5 𝝁m 1.749527 0.620239

DPVuit6 Mean Standard Deviation

Exhaust 0.3-0.5 𝝁m 5295.169 292.953 Exhaust >0.5 𝝁m 17.87271 8.14789 Outside 0.3-0.5 𝝁m 138.5149 54.43331 Outside >0.5 𝝁m 1.280857 0.474464

EPVaan4 Mean Standard Deviation


EPVuit3 Mean Standard Deviation


FPVaan3 Mean Standard Deviation


FPVuit4 Mean Standard Deviation




Appendix J: Used Matlab M-files This appendix shows the program files which are used in Matlab to post process the data. M_file reading data of CO2 measurements This file is used in the static experiments and the continuous experiments. This file translates data and combines data of the B&K with data of the squirrellogger. clear all close all clc %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% load B1_bk %% #1 data_set = B1_bk; %% achter = teken -> tik naam CSV file #2 load B1_sq %% #3 data_set_sq = B1_sq; %% achter = teken tik naam CSV file #4 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% BK uitlezen data_ch1 = []; data_ch2 = []; data_ch3 = []; data_ch4 = []; data_ch5 = []; for i = 1:length(data_set) if data_set(i,1)==1 data_ch1 = [data_ch1; data_set(i,1) data_set(i,2) data_set(i,3)]; elseif data_set(i,1)==2 data_ch2 = [data_ch2; data_set(i,1) data_set(i,2) data_set(i,3)]; elseif data_set(i,1)==3 data_ch3 = [data_ch3; data_set(i,1) data_set(i,2) data_set(i,3)]; elseif data_set(i,1)==4 data_ch4 = [data_ch4; data_set(i,1) data_set(i,2) data_set(i,3)]; elseif data_set(i,1)==5 data_ch5 = [data_ch5; data_set(i,1) data_set(i,2) data_set(i,3)]; end end %%%%%Squirrel koppellen aan BK CH1_bk_sq = []; for i = 1:length(data_ch1) tijd = data_ch1(i,2); L = find(tijd<=data_set_sq(:,1)); data_sq_klein_ch1(i,:) = data_set_sq(L(1),:); data_totaal_ch1(i,:) = [data_ch1(i,:), data_sq_klein_ch1(i,:)]; end CH2_bk_sq = []; for i = 1:length(data_ch2) tijd = data_ch2(i,2); L = find(tijd<=data_set_sq(:,1)); data_sq_klein_ch2(i,:) = data_set_sq(L(1),:); data_totaal_ch2(i,:) = [data_ch2(i,:), data_sq_klein_ch2(i,:)]; end Etcetera.... %%%% Omzetten voltage co2 naar ppm data_ch1_328=-1.5835*data_totaal_ch1(:,5).*data_totaal_ch1(:,5)+192.3676*data_totaal_ch1(:,5)+76.7928; data_ch1_341=-1.5373*data_totaal_ch1(:,6).*data_totaal_ch1(:,6)+191.6807*data_totaal_ch1(:,6)+66.4161; data_ch1_329=(431.66*data_totaal_ch1(:,7)-125.53); data_totaal_ch1=[data_totaal_ch1(:,1:4),data_ch1_328,data_ch1_341,data_ch1_329,data_totaal_ch1(:,8:10)]; data_ch2_328=-1.5835*data_totaal_ch2(:,5).*data_totaal_ch2(:,5)+192.3676*data_totaal_ch2(:,5)+76.7928; data_ch2_341=-1.5373*data_totaal_ch2(:,6).*data_totaal_ch2(:,6)+191.6807*data_totaal_ch2(:,6)+66.4161; etcetera.... %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% B1_ch1 = data_totaal_ch1; B1_ch2 = data_totaal_ch2; B1_ch3 = data_totaal_ch3; B1_ch4 = data_totaal_ch4; B1_ch5 = data_totaal_ch5; save B1_ch1 B1_ch1 save B1_ch2 B1_ch2 save B1_ch3 B1_ch3 save B1_ch4 B1_ch4 save B1_ch5 B1_ch5

M-File Statistics



This file is used in the static experiments and the continuous experiments and calculates the mean, mode and standard deviation of all values. Also an algorithm is opened to test the normal distribution. % met ctrl f kan veranderd worden. Oppassen met B1 onderste regel % uitvoer is een matrix statistiek_xx waarvan de rijen de verschillende % kanalen 1 t/m 5 zijn en de kolommen respectievelijk % 1. Gemiddelde B&K % 2. Standaarddeviatie B&K % 3. Gemiddelde Toevoerlucht % 4. Standaarddeviatie toevoerlucht % 5. Gemiddelde Afvoerlucht % 6. Standaarddeviatie Afvoerlucht % 7. Gemiddelde controlepunt % 8. Standaarddeviatie controlepunt. clear all close all clc load B1_ch1 load B1_ch2 load B1_ch3 load B1_ch4 load B1_ch5 %%%% statistiek mean_bk(1,1)=mean(B1_ch1(:,3)); mean_bk(2,1)=mean(B1_ch2(:,3)); mean_bk(3,1)=mean(B1_ch3(:,3)); mean_bk(4,1)=mean(B1_ch4(:,3)); mean_bk(5,1)=mean(B1_ch5(:,3)); median_bk(1,1)=median(B1_ch1(:,3)); median_bk(2,1)=median(B1_ch2(:,3)); median_bk(3,1)=median(B1_ch3(:,3)); median_bk(4,1)=median(B1_ch4(:,3)); median_bk(5,1)=median(B1_ch5(:,3)); mode_bk(1,1)=mode(B1_ch1(:,3)); mode_bk(2,1)=mode(B1_ch2(:,3)); mode_bk(3,1)=mode(B1_ch3(:,3)); mode_bk(4,1)=mode(B1_ch4(:,3)); mode_bk(5,1)=mode(B1_ch5(:,3)); [h_lillie(1,1),p_lillie(1,1)]=lillietest(B1_ch1(:,3)); [h_lillie(2,1),p_lillie(2,1)]=lillietest(B1_ch2(:,3)); [h_lillie(3,1),p_lillie(3,1)]=lillietest(B1_ch3(:,3)); [h_lillie(4,1),p_lillie(4,1)]=lillietest(B1_ch4(:,3)); [h_lillie(5,1),p_lillie(5,1)]=lillietest(B1_ch5(:,3)); mean_supply_air(1,1)=mean(B1_ch1(:,5)); mean_supply_air(2,1)=mean(B1_ch2(:,5)); mean_supply_air(3,1)=mean(B1_ch3(:,5)); mean_supply_air(4,1)=mean(B1_ch4(:,5)); mean_supply_air(5,1)=mean(B1_ch5(:,5)); mean_exhaust_air(1,1)=mean(B1_ch1(:,6)); mean_exhaust_air(2,1)=mean(B1_ch2(:,6)); mean_exhaust_air(3,1)=mean(B1_ch3(:,6)); mean_exhaust_air(4,1)=mean(B1_ch4(:,6)); mean_exhaust_air(5,1)=mean(B1_ch5(:,6)); mean_control_point(1,1)=mean(B1_ch1(:,7)); mean_control_point(2,1)=0; mean_control_point(3,1)=0; mean_control_point(4,1)=0; mean_control_point(5,1)=0; std_bk(1,1)=std(B1_ch1(:,3)); std_bk(2,1)=std(B1_ch2(:,3)); std_bk(3,1)=std(B1_ch3(:,3)); std_bk(4,1)=std(B1_ch4(:,3)); std_bk(5,1)=std(B1_ch5(:,3)); std_supply_air(1,1)=std(B1_ch1(:,5)); std_supply_air(2,1)=std(B1_ch2(:,5)); std_supply_air(3,1)=std(B1_ch3(:,5)); std_supply_air(4,1)=std(B1_ch4(:,5)); std_supply_air(5,1)=std(B1_ch5(:,5)); std_exhaust_air(1,1)=std(B1_ch1(:,6)); std_exhaust_air(2,1)=std(B1_ch2(:,6)); std_exhaust_air(3,1)=std(B1_ch3(:,6)); std_exhaust_air(4,1)=std(B1_ch4(:,6)); std_exhaust_air(5,1)=std(B1_ch5(:,6)); std_control_point(1,1)=std(B1_ch1(:,7)); std_control_point(2,1)=0; std_control_point(3,1)=0; std_control_point(4,1)=0; std_control_point(5,1)=0; statistiek_B1=[mean_bk,std_bk,median_bk,mode_bk,h_lillie,p_lillie,mean_supply_air,std_supply_air,mean_exhaust_air,std_exhaust_air,mean_control_point,std_control_point]; save statistiek_B1 statistiek_B1 xlswrite('Statistieken_meetdata.xlsx',statistiek_B1,'B1','B4');



M-File Effectiveness and local air quality This file is used in the static experiments and the continuous experiments and calculates the effectiveness and local air quality of all cases. %Berekend de Effictiviteit, met ctrl F kan zowel de index (bijv B2) als de %referentie(bijv B1) veranderd worden clear all close all clc load statistiek_B1 load statistiek_B2 %eff= ventilation index at positition of each channel: %effPV_xxtoyy= efficienty van xx tov yy for i = 1:5 %eerst bekijken of B&K data normaal verdeeld is met lilliefors test. %altijd mediaan %if statistiek_B2(i,5)==0; % waarde_B2(i,1)=statistiek_B2(i,1); %elseif statistiek_B2(i,5)==1; waarde_B2(i,1)=statistiek_B2(i,3)+2*statistiek_B2(i,2); % end % if statistiek_B1(i,5)==0; % waarde_B1(i,1)=statistiek_B1(i,1); %elseif statistiek_B1(i,5)==1; waarde_B1(i,1)=statistiek_B1(i,3)-2*statistiek_B1(i,2); % end localairchangeindex_B2(i,1)=statistiek_B2(i,9)/waarde_B2(i,1); localairchangeindex_B1(i,1)=statistiek_B1(i,9)/waarde_B1(i,1); datapv(i,1)=((statistiek_B2(i,9)-statistiek_B2(i,7))/(waarde_B2(i,1)-statistiek_B2(i,7))); datanul(i,1)=((statistiek_B1(i,9)-statistiek_B1(i,7))/(waarde_B1(i,1)-statistiek_B1(i,7))); effPV_B2toB1(i,1)=(datapv(i,1)-datanul(i,1))/datapv(i,1); end %% print waardes localairchangeindex_B1 localairchangeindex_B2 effPV_B2toB1

M_file Plot CO2 data This file is used in the static experiments and the continuous experiments and plots all data in one graph. %%%% met control F kunnen de verschillende versies veranderd worden %% deze M-file maakt een overzichtplaatje voor alle kanalen samen. clear all close all clc load C4_ch1 load C4_ch2 load C4_ch3 load C4_ch4 load C4_ch5 %full screen figure screen_size = get(0, 'ScreenSize'); f1 = figure(1); set(f1, 'Position', [0 0 screen_size(3) screen_size(4) ] ) %%%%%%%%%%_ch1 time_ch1=(C4_ch1(:,2)-C4_ch1(1,2))*1440; subplot(2,4,1) ph1=plot(time_ch1,C4_ch1(:,3)); hold on ph2=plot(time_ch1,C4_ch1(:,5),'g'); hold on ph3=plot(time_ch1,C4_ch1(:,6),'r'); hold on %ph4=plot(time_ch1,C4_ch1(:,7),'c'); %hold on ylabel('ppm CO2') xlabel('time [min]') Title('C4 ch1 CO2 concentration') subplot(2,4,7) ph5=plot(time_ch1,C4_ch1(:,8),'c:'); hold on ph6=plot(time_ch1,C4_ch1(:,9),'g:'); hold on ph7=plot(time_ch1,C4_ch1(:,10),'m:'); ylabel('Temperature [C]') xlabel('time [min]') Title('C4 temperature') %legend('Room','Sup','Out','Location','EastOutside') %%%%%%%%%%_ch2 time_ch2=(C4_ch2(:,2)-C4_ch2(1,2))*1440;



subplot(2,4,2) plot(time_ch2,C4_ch2(:,3)) hold on plot(time_ch2,C4_ch2(:,5),'g') hold on plot(time_ch2,C4_ch2(:,6),'r') hold on %plot(time_ch2,C4_ch2(:,7),'c') hold on ylabel('ppm CO2') xlabel('time [min]') Title('C4 ch2 CO2 concentration') %%%%%%%%%%_ch3 time_ch3=(C4_ch3(:,2)-C4_ch3(1,2))*1440; subplot(2,4,3) plot(time_ch3,C4_ch3(:,3)) hold on plot(time_ch3,C4_ch3(:,5),'g') hold on plot(time_ch3,C4_ch3(:,6),'r') hold on %plot(time_ch3,C4_ch3(:,7),'c') hold on ylabel('ppm CO2') xlabel('time [min]') Title('C4 ch3 CO2 concentration') %legend('B&K','Sup','Exh','Location','EastOutside') %%%%%%%%%%_ch4 time_ch4=(C4_ch4(:,2)-C4_ch4(1,2))*1440; subplot(2,4,5) plot(time_ch4,C4_ch4(:,3)) hold on plot(time_ch4,C4_ch4(:,5),'g') hold on plot(time_ch4,C4_ch4(:,6),'r') hold on %plot(time_ch4,C4_ch4(:,7),'c') hold on ylabel('ppm CO2') xlabel('time [min]') Title('C4 ch4 CO2 concentration') %%%%%%%%%%_ch5 time_ch5=(C4_ch5(:,2)-C4_ch5(1,2))*1440; subplot(2,4,6) plot(time_ch5,C4_ch5(:,3)) hold on plot(time_ch5,C4_ch5(:,5),'g') hold on plot(time_ch5,C4_ch5(:,6),'r') hold on %plot(time_ch5,C4_ch5(:,7),'c') hold on ylabel('ppm CO2') xlabel('time [min]') Title('C4 ch5 CO2 concentration') sh=subplot(2,4,4); p=get(sh,'position'); lh=legend(sh,[ph1,ph2,ph3,ph5,ph6,ph7],'B&K CO2','Supply air CO2','Exhaust air CO2','Room Temp','Supply air Temp','Outside temp'); set(lh,'position',p); set(lh,'Fontsize',6); axis(sh,'off')

M_file particle counts File used for the single movement measurements This M-file translates voltages to particle counts, plots the graph and boundary condition graph and thereafter calculates the integral of the graph every 10seconds after the movement has started and plots the results in bar plot. clear all clc close all load EPVaan2 EPVaan2



max03=5000; %instelling sensor max05=1000; %instelling sensor Referentie03=0; Referentie05=0; checktime=30; %time after movements Deeltjes03=[]; Deeltjes05=[]; Tijd=[]; Movement=[]; for i=1:length(EPVaan2) if EPVaan2(i,6) <=0.8 Deeltjes03(i,:)=max03/4; elseif EPVaan2(i,6) >=0.8 Deeltjes03(i,:)=(EPVaan2(i,6)-0.94).*(max03/4/4); % eerste 4 = volt, 4 =period end if EPVaan2(i,7) <=0.8 Deeltjes05(i,:)=max05/4; elseif EPVaan2(i,7) >=0.8 Deeltjes05(i,:)=(EPVaan2(i,7)-0.94).*(max05/4/4); end if EPVaan2(i,8) <=0.8 Outside03(i,:)=(6000/30)*2.01; elseif EPVaan2(i,8) >=0.8 Outside03(i,:)=((EPVaan2(i,8)-0.94).*(6000/4/30))*2.01; end if EPVaan2(i,9) <=0.8 Outside05(i,:)=(1500/30)*0.6; elseif EPVaan2(i,9) >=0.8 Outside05(i,:)=((EPVaan2(i,9)-0.94).*(1500/4/30))*0.6; end if EPVaan2(i,10) <=0.8 Exhaust03(i,:)=(400000/60)*1.95; elseif EPVaan2(i,10) >=0.8 Exhaust03(i,:)=((EPVaan2(i,10)-0.94).*(400000/4/60))*1.95; end if EPVaan2(i,11) <=0.8 Exhaust05(i,:)=(12000/60)*0.26; elseif EPVaan2(i,11) >=0.8 Exhaust05(i,:)=((EPVaan2(i,11)-0.94).*(12000/4/60))*0.26; end Tijd(i,:)=(EPVaan2(i,1)-EPVaan2(1,1))*24*60; if EPVaan2(i,5)>=0.1 Movement(i,:)=0; elseif EPVaan2(i,5)==0 Movement(i,:)=max03/4; end end x=1:1:length(EPVaan2); Tijds=Tijd*60; %%%Plotten van figuur figure plot(Tijds,Deeltjes03) hold on plot(Tijds,Deeltjes05,'r') plot(Tijds,Movement,'g') %plot(Tijds,fit03,'c') title('EPVaan2') xlabel('time [s]') ylabel('particle counts per s') legend('Particles between 0.3 and 0.5 microm','Particles larger than 0.5 microm','Movement') figure subplot(2,2,1) ph1=plot(Tijds,Outside03,''); hold on ph2=plot(Tijds,Outside05,'r'); title('Outside particles') xlabel('time [s]') ylabel('particle counts per s') subplot(2,2,2) ph3=plot(Tijds,Exhaust03); hold on ph4=plot(Tijds,Exhaust05,'r'); title('Exhaust particles') xlabel('time [s]') ylabel('particle counts per s') subplot(2,2,3) ph5=plot(Tijds,EPVaan2(:,2),'g') hold on ph6=plot(Tijds,EPVaan2(:,3),'c') hold on



ph7=plot(Tijds,EPVaan2(:,4),'m') title('Temperature') xlabel('time [s]') ylabel('Temperature [C]') sh=subplot(2,2,4); p=get(sh,'position'); lh=legend(sh,[ph1,ph2,ph5,ph6,ph7],'Particles between 0.3 and 0.5 microm','Particles larger than 0.5 microm','Room Temp','Supply air Temp','Outside temp'); set(lh,'position',p); set(lh,'Fontsize',6); axis(sh,'off') mtit('EPVaan2','fontsize',12,'color',[1 0 0]) %grid %%Array invoeren die de start van de beweging registreert Changingmovement=[]; for i=5:length(Movement); if Movement(i,1)>0 && sum(Movement(i-1:i-1,1))==0; Changingmovement(i,1)=Movement(i,1); elseif Movement(i,1)==0 || sum(Movement(i-1:i-1,1))>0; Changingmovement(i,1)=0; end end aantalbew=sum((Changingmovement(:,1)>0)); %% Pieken wegschrijven naar een gedeelte wat random is %% en een gedeelte wat vlak na de beweging plaastvind. Tabel_EPVaan2=[Tijd,Tijds,Deeltjes03,Deeltjes05,Changingmovement]; Tabel2_EPVaan2=[]; for i=checktime+1:length(Tabel_EPVaan2); if Referentie03<Tabel_EPVaan2(i,3) && sum(Tabel_EPVaan2((i-checktime):i,5))>1; Tabel2_EPVaan2(i,1)=Tabel_EPVaan2(i,3)-Referentie03; elseif Referentie03>Tabel_EPVaan2(i,3) && sum(Tabel_EPVaan2(i-checktime:i,5))>1; Tabel2_EPVaan2(i,1)=0; elseif Referentie03<Tabel_EPVaan2(i,3) && sum(Tabel_EPVaan2(i-checktime:i,5))<1; Tabel2_EPVaan2(i,2)=Tabel_EPVaan2(i,3)-Referentie03; elseif Referentie03>Tabel_EPVaan2(i,3) && sum(Tabel_EPVaan2(i-checktime:i,5))<1; Tabel2_EPVaan2(i,2)=0; end end for i=checktime+1:length(Tabel_EPVaan2); if Referentie05<Tabel_EPVaan2(i,4) && sum(Tabel_EPVaan2((i-checktime):i,5))>1; Tabel2_EPVaan2(i,3)=Tabel_EPVaan2(i,4)-Referentie05; elseif Referentie05>Tabel_EPVaan2(i,4) && sum(Tabel_EPVaan2(i-checktime:i,5))>1; Tabel2_EPVaan2(i,3)=0; elseif Referentie05<Tabel_EPVaan2(i,4) && sum(Tabel_EPVaan2(i-checktime:i,5))<1; Tabel2_EPVaan2(i,4)=Tabel_EPVaan2(i,4)-Referentie05; elseif Referentie05>Tabel_EPVaan2(i,4) && sum(Tabel_EPVaan2(i-checktime:i,5))<1; Tabel2_EPVaan2(i,4)=0; end end %%verdelen per 10sec na beweging for i=10+1:length(Tabel_EPVaan2); if Referentie03<Tabel_EPVaan2(i,3) && sum(Tabel_EPVaan2((i-10):i,5))>1; Tabel2_EPVaan2(i,5)=Tabel_EPVaan2(i,3)-Referentie03; elseif Referentie03>Tabel_EPVaan2(i,3) && sum(Tabel_EPVaan2(i-10:i,5))>1; Tabel2_EPVaan2(i,5)=0; end end for i=20+1:length(Tabel_EPVaan2); if Referentie03<Tabel_EPVaan2(i,3) && sum(Tabel_EPVaan2((i-20):i,5))>1; Tabel2_EPVaan2(i,6)=Tabel_EPVaan2(i,3)-Referentie03; elseif Referentie03>Tabel_EPVaan2(i,3) && sum(Tabel_EPVaan2(i-20:i,5))>1; Tabel2_EPVaan2(i,6)=0; end end ETC. %%Integralen van de pieken uitrekenen. int03duetomovements=sum(Tabel2_EPVaan2(:,1)); int03random=sum(Tabel2_EPVaan2(:,2)); int05duetomovements=sum(Tabel2_EPVaan2(:,3)); int05random=sum(Tabel2_EPVaan2(:,4)); Totaal03=sum(Deeltjes03(:,1)); Totaal05=sum(Deeltjes05(:,1)); Gem03per10s=Totaal03/(length(EPVaan2)/(10)); Gem05per10s=Totaal05/(length(EPVaan2)/(10)); Gemgrafiek03=[Gem03per10s,Gem03per10s,Gem03per10s,Gem03per10s,Gem03per10s,Gem03per10s,Gem03per10s,Gem03per10s,Gem03per10s,Gem03per10s,Gem03per10s,Gem03per10s]; Gemgrafiek05=[Gem05per10s,Gem05per10s,Gem05per10s,Gem05per10s,Gem05per10s,Gem05per10s,Gem05per10s,Gem05per10s,Gem05per10s,Gem05per10s,Gem05per10s,Gem05per10s]; cumint03mov10=sum(Tabel2_EPVaan2(:,5)); cumint03mov20=sum(Tabel2_EPVaan2(:,6)); cumint03mov30=sum(Tabel2_EPVaan2(:,7));



ETC. int03mov10=(cumint03mov10)/aantalbew; int03mov20=(cumint03mov20-cumint03mov10)/aantalbew; int03mov30=(cumint03mov30-cumint03mov20)/aantalbew; Etc. int05mov10=(cumint05mov10)/aantalbew; int05mov20=(cumint05mov20-cumint05mov10)/aantalbew; int05mov30=(cumint05mov30-cumint05mov20)/aantalbew; Etc. gegevens=[Referentie03; Referentie05; int03duetomovements; int03random; int05duetomovements; int05random; Totaal03; Totaal05; Gem03per10s;Gem05per10s;aantalbew]; gegevens10cum=[cumint03mov10;cumint03mov20;cumint03mov30;cumint03mov40;cumint03mov50;cumint03mov60;cumint03mov70;cumint03mov80;cumint03mov90;cumint03mov100;cumint03mov110;cumint03mov120 cumint05mov10;cumint05mov20;cumint05mov30;cumint05mov40;cumint05mov50;cumint05mov60;cumint05mov70;cumint05mov80;cumint05mov90;cumint05mov100;cumint05mov110;cumint05mov120]; gegevens10=[int03mov10;int03mov20;int03mov30;int03mov40;int03mov50;int03mov60;int03mov70;int03mov80;int03mov90;int03mov100;int03mov110;int03mov120 int05mov10;int05mov20;int05mov30;int05mov40;int05mov50;int05mov60;int05mov70;int05mov80;int05mov90;int05mov100;int05mov110;int05mov120]; xlswrite('analyse_deeltjes.xlsx',gegevens,'EPVaan2','B3'); xlswrite('analyse_deeltjes.xlsx',gegevens10cum,'EPVaan2','D3'); xlswrite('analyse_deeltjes.xlsx',gegevens10,'EPVaan2','E3'); gegevens10_EPVaan2=[]; gegevens10_EPVaan2=[gegevens10;Gem03per10s;Gem05per10s;aantalbew;length(EPVaan2)]; save gegevens10_EPVaan2 gegevens10_EPVaan2 figure x=10:10:120; labelsxas=; bar(x,gegevens10(1:12,1)); hold on plot(x,Gemgrafiek03) hold on bar(x,gegevens10(13:24,1),0.4,'r'); hold on plot(x,Gemgrafiek05,'r') ylabel('Number of particles') xlabel('Time after movement [s]') set(gca, 'XTickLabel', labelsxas) legend('Particles size between 0.3 and 0.5 microm','Average concentration 0.3-0.5 microm','Particles size > 0.5 microm','Average concentration > 0.5 microm')

Personal ventilation to control airborne infectious diseases in … · To control infection spread,...

Documents

Transcript of Personal ventilation to control airborne infectious diseases in … · To control infection spread,...