Text Size
Friday, July 28, 2017

Website started and updated by faculty and students of Cal Poly Pomona's College of Environmental Design: Department of Architecture.

"A sustainable online forum for students

and young professionals of design."

Blog Content

Thermal Comfort Simulations for Disaster-Relief Housing for Haiti (ASES 2012)

This paper is from the ASES Proceedings - which can viewed with images here or at the ASES website.


This paper summarizes a methodological approach for incorporating sustainable principles and the appropriate use of simulation tools into the design process. The work was completed in a year and a half of cumulatively coordinated courses taught at California Poly Pomona. Architecture students learned to use simulation tools to optimize the design for optimal indoor temperature, daylighting levels, and airflow. Throughout the process, environmental simulations and consequent design changes were carried out, informing one another in an iterative, integrated design process. The students designed low-income housing and other buildings for disaster-stricken Haiti using a unique honeycomb panel system made from recycled paper that is prefabricated and assembled in Haiti exclusively with Haitian labor. Methodologically, students learned an environmental simulation workflow to assess the design for human comfort. The primary goal of the simulation efforts was to make appropriate and timely improvements to the designs for thermal comfort where cost and simplicity are primary constraints.


Approximately two years after the magnitude-7 earthquake in Haiti in January of 2010, more than 500,000 Haitians still remain homeless, living in camps with little more than tents or tarps over their heads1. The housing project hinges on enabling Haitians through being built entirely in Haiti, with local labor and a minimum of imported materials (Fig. 1) (Fig.2). In addition to the considerations relevant to disaster relief housing such as cultural context, affordability, and structural stability, this project focuses on design optimized for passive strategies to improve the comfort levels of occupants within these structures. The entire project was realized through a series of courses at Cal Poly Pomona in partnership with a Haitian factory. The implementation spanned over an entire academic year with a series of cumulative courses focused on design, simulation and analysis.

Although this paper is focused on a methodological approach based on simulation tools, it is important to point out that the process was very much influenced by a number of issues inherent in such a real-world disaster relief project. From the onset, the project was tied to a Haitian company, Haitian Building Solutions, who was intent on working with a building product for housing which could simultaneously enable local labor and minimize the need for imported materials (Fig. 2). Students therefore initially had to learn to design with a very specific modular building material. Central to our constraints was to design with this unique honeycomb composite panel which could be manufactured locally through a factory established in Haiti. Further, the design, manufacturing process and construction of the housing had to be carried out within an informed and sensitive understanding of the cultural context of Haiti. Consequent to the use of this experimental panel, students also focused on the actual prototyping of details and a full-scale mockup of the structure of the housing.

1.1.  Local Empowerment

Heavily influencing the design as well was the fact that the central goal of our Haitian manufacturing facility, Haiti Building solutions was job creation in Haiti. The unemployment estimates vary widely but estimates place it at around 70-90% 2. This project is uniquely constrained in its efforts to exclusively use both local labor and local materials. The houses are constructed almost solely with a unique resin-coated corrugated paper core sandwiched between and glued to magnesium board panels which are manufactured locally (Fig. 3). The panels are also simple to assemble from a construction standpoint, and a fully functional move-in ready house can be assembled by five persons in as little as four days. The cost of a 53 square meter finished home (utilizing structural panels for floor) is $10,000 USD, or approximately $20/sf.

The advantages of the honeycomb composite panel include 1) a great reduction of the need for importing materials, 2) creating lightweight housing that will be safer in the event of structural failure, 3) a very fast erection time, and 4) the integration of local labor. The panels are impervious to fire, water, insects, mold, and mildew. In addition, efforts are made to source nearly all of the additional building materials locally. 

2.       Design Process

Designs were carried out using this panelized system during an architectural studio course and were presented at the Building Back Better Communities (BBBC) Expo in February of 2011.Of the designs presented, one house, the Takit-EZ House (Fig. 4), was selected to be constructed at the Expo site. The Takit-EZ house was designed by students and built both as a prototype on campus and as a fully functional house in Port-au-Prince Haiti as part of the BBBC Expo organized by the Haitian government. This expo showcased model homes by approved vendors, including HBS. This particular house was the only house of sixty model homes built entirely in Haiti, with Haitian labor (see Figure 2). The design was chosen because of its asymmetric roofline, efficient plan, expandability, clerestory windows and French doors which encouraged natural ventilation and connections to the outdoor areas. The enlarged porch allowed for outdoor living, as the heat indoors is unbearable much of the year in Haiti. This paper describes the subsequent design and simulation work done to improve the design of this project.

2.1.  Haitian Climate

In Haiti, the temperature is hot and humid, with temperatures that ranging from on average from 73° F to 88° F (23-31° C) in winter and 77–95° F (25-35° C) in summer3 (Fig. 5). The wet season is from April-June and October-November, and Haiti is subject to frequent hurricanes, with wind speeds up to 150 mph. Port-au-Prince is located at 18º 32' N and 72º 20' W. This is south of the Tropic of Cancer, and during summer months, the sun is frequently to the north of the building (Fig. 6). In winter and spring, the sun is primarily to the south. Winds blow predominantly from ENE at an average speed of 8mph (3.6 mps) the site (Fig. 7).

2.2.  Passive Strategies

The design incorporates a number of passive strategies in order to increase the thermal comfort of its inhabitants. The primary strategies for this climate type are natural ventilation, shading, and the use of high-albedo surfaces. The project was designed to encourage cross ventilation and stack ventilation. Each inhabitable room has windows on two different sides to encourage airflow. The main room has doors that open up in front and back as well as clerestory windows to allow hot air to escape. For shading, exterior devices in the form of shutters were designed. These shutters act to keep the sun and rain out when needed. Additionally, the roof is painted white and the walls bright colors in order to reflect solar heat gain.

As mentioned, the designs were carried out in a design studio during winter 2011 and subsequently optimized and further developed in a seminar taught during winter 2012. Due to cost concerns, there was not much freedom with respect to shading design which was dictated by panel sizing. Also, window placement was heavily affected by shear wall locations. The areas where design could be changed included building geometry, window placement, local shading strategies, and insulation. These variations were tested during the seminar course and are discussed below.

2.3.  Ventilation and Orientation

The first step in optimization was to adapt the design for the varying possibilities of orientations. The original design did not consider orientation, as it was a prototypical design that would be used in all possible orientations. The goal was to make specific changes based on orientation to the design that were easily carried out by the factory, for example mirroring the whole design, but not changes that required new details for construction.

The west wall is subject to the strongest solar radiation values during the summer (Fig.8). Therefore for this orientation, the bedrooms have been up put on the east side, as inhabitants spend a significant time in the bedrooms. Airflow was analyzed for this orientation, and it was found that the airflow through the bedrooms is greatly improved when the bedrooms are on the east side (Fig. 9) (Fig.10). A further reason for the mirroring of the geometry is so that the clerestory windows face the leeward side, so that hot air can be encouraged to flow up and out of these windows.

The north facing orientation makes use the original design geometry. In this case, the bedrooms face east and the clerestory windows again point leeward. Wind Tunnel studies show that the airflow is good through all rooms except the bathroom, which is not critical (Fig. 11). Several variations of this orientation were tested, including night where the front doors are closed (Fig. 12) and a version where all doors are closed.

For the west-facing condition, the building geometry is the same as the north orientation. This provides for bedrooms at the north, on the windward side. The rationale is that the airflow is more critical in the bedrooms due to the fact that Haitians often spend a great deal of time during the day outdoors. Once again, the clerestory windows are at the leeward side. For the east-facing orientation, the building geometry mimics that of the south orientation, so that bedrooms are located at the north, and clerestory window is located on the leeward side.

2.4.  Analysis of Base Condition

The designs were then analyzed for thermal comfort. Initially, they were modeled and analyzed in DesignBuilder (Fig. 13). The goal was to improve the design where changes were possible. The parameters of the walls were set as magnesium board facings with an air gap between boards to represent the honeycomb core, and a layer of cement plaster on the outside. The roof and floor were similarly set, except the air gap was thicker, and tile was added to the floor. No insulation was included.

The floor was raised above ground. Windows were set to have the lowest R value allowed by DesignBuilder. Natural ventilation was scheduled to always be on, and all mechanical and lighting systems were turned off. It was assumed that all the windows (and doors) were open all the time. The house was divided into 3 blocks: the bedrooms, the kitchen/bath wing, and the living area. These blocks were further divided into the following zones: northeast bedroom, southeast bedroom, kitchen, bath, and living area.

Firstly the north-facing design (with east bedrooms) was tested to see the effects of natural ventilation  for understanding the impact of our most effective passive strategy (Fig. 14). The cooling design was run for the summer design day. Natural ventilation proved to have a big impact as peak temperature reduced from 112° F (44° C ) in the late afternoon on the peak summer day to 98° F (37° C ). These temperatures are still too high for thermal comfort, and subsequent studies will work to get these lower. A further analysis of the individual zones indicated the following that the bedrooms are the hottest of all the areas (Fig. 15).

These values were also compared to the original design, with the buildings rotated to face south and bedrooms to the west (Fig.16). The bedrooms are still very hot, and in fact much hotter in the original orientation (when they are facing west). The peak temperature for the public spaces did not change by as much, but the peak time changed from 3PM to 2PM. In the case of the bedrooms, the peak time shifted from 2PM to 5PM (for the original version), and temperature reduced from 105° F (41° C ) to 101° F (33° C ). It should be mentioned that this project was approached by looking at the peak temperatures on the peak day in an effort to reduce the most uncomfortable period. As shown in the table below and in Fig. 17, the bedrooms are greatly affected by the change from west-facing bedrooms to east-facing bedrooms.



Bedrooms face east

Bedrooms face west


Max. Temp.° F

Min. Temp.° F

Max. Temp.° F

Min. Temp.° F


NE Bedroom






SE Bedroom

























2.5.  Design Optimization

In looking at the graph of fabric gains, much of the heat gain is coming from the roof. In an effort to reduce these gains, a cooling design simulation was run with the addition of insulation in the roof. Foam insulation was inserted into the layer of honeycomb, so it was approximately 6" thick (0.15m). The results were dramatic (Fig. 18) (Fig.19), with the curved shape of the roof heat gain disappearing from view in the chart.

Next the effect of insulating walls, partitions, and floors were tested individually. The walls and partitions are 3.5" thick (0.1m) and also received insulation between the magnesium board, and the floor was similar in thickness to the roof. The results show minor reduction of heat gain during the day, but it is interesting to note that the insulation also seems to impede nighttime cooling by a slight amount (Fig. 20). The graph shown in Fig. 21 indicates improvements to indoor temperature as a function of insulation, with improvements as the insulation is increased. Insulation for the roof provides the most significant benefit, with reductions of approximately 3° F (1.5° C). Insulating all surfaces reduces peak temperature by 7° F (4° C) from 100° F (38° C) to 93°F (34° C).

Next, the effects of shading were tested. 0.5m fins were added to all openings, and internal blinds were also added. Overall, this did not make a very significant change. Next, the floor was changed from a raised floor to slab on grade, and this was also combined with the roof insulation. The changes were significant. Fig. 20 shows the effect of insulating the roof and using slab on grade construction provide temperatures close to outside air temperature (Fig. 22).

2.6.  Daylighting

Next the house was studied to evaluate the indoor light levels. The houses have no electricity, and rely exclusively on natural sunlight as the only source of light. The worst case and intermediate scenarios were tested in Ecotect and Radiance, and the results met the minimum daylighting levels for each of the respective spaces (Fig. 23) (Fig. 24) (Fig. 25).

3.       summary and FUTURE STEPS

The simulations indicated that the best results included: designing for orientation so that the bedrooms were at the east, the roof was insulated, and the floor was changed to slab on grade. In this condition, the indoor temperatures came close to outdoor temperatures, a desirable condition when there is no HVAC system present. However, these strategies may be too expensive, so it would be preferable to continue to investigate the design so that these lowered temperature conditions can be achieved with less expensive techniques. This paper focuses on the environmental simulations, but it should be mentioned that the prototyping and construction process served as an equally valuable tool. Students gained insight from experiences that only come about when doing full-scale construction (Fig. 26).

3.1.  Expanded Scope

It is hoped that future projects will test a larger variety of window and shading conditions, including the effect of schedules. In addition, we plan to analyze larger houses that use HVAC and renewable systems. This would be a complementary skill- building exercise for students as the output of heating and/or cooling loads and percent savings may be easier to analyze. In addition, several clients of HBS are seeking a middle-class solution to the housing crisis. In these circumstances, the goal would be to minimize any cooling system utilized and to supply the energy needed entirely with renewable resources.

3.2.  Measurement and Verification

This project also hopes to test the validity of the tools by measuring environmental conditions of built projects in Haiti and comparing these to simulation values. The most likely projects to be analyzed are the Takit-EZ House (built), and an Orphanage which is currently under construction. The orphanage houses 30 children, and the intent is to insert an entire set of sensors to measure indoor and outdoor environmental conditions, including temperature, and humidity, in addition to taking surveys to question occupant comfort levels. These values will in turn be compared to the simulation results for the same. In addition, airflow and light will be measured and compared to values form Vasari and Ecotect/Radiance. This would be a valuable step providing verification of the tools used.

4.       acknowledgements

The following students took part in the classes mentioned in this paper.

The following students took part in the projects mentioned in this paper: Design Team - Liliana Alvarez, Gabriela Barajas, Nicole Graciano, Jennifer Guerra, Ashi Martin, Fariba Mostajer, Ryan Raskop,  Build Team - Maro Asipyan, Peter Fox, Nathan Houck, Ryan Raskop, Miguel Simental, Christopher Stanford, Matthew Terry, Caleb Wong, and Christopher Young.

The author would also like to thank the following instructors for providing generously with their time and expertise: Michael Fox, Gary McGavin, Mikhail Gershfeld, and Erik Mar.

5.       References

 (1) "Haitian authorities forcing quake victims into homelessness." Amnesty International. 11 Jan 2012. 10 March 2012. http://www.amnesty.org/en/news/haitian-authorities-forcing-quake-victims-homelessness-2012-01-11 (2) "Dumas: Unemployment in Haiti now 90 per cent former UN adviser calls for investment." Guardian Media. 13 Jan 2012. 10 March 2012. http://www.guardian.co.tt/ news/friday-january-13-2012/dumas-unemployment-haiti-now-90-cent-former-un-adviser-calls-investment (3) "Haiti - Climate." Encyclopedia of the Nations. 10 March 2012. http://www.nationsencyclopedia.com/Americas/Haiti-CLIMATE.html#b